2025Activity​ reportProject-TeamGENSCALE

3.1.1​​ Representation & Compression

Participants:​​​‌ Karel Brinda, Pierre‌ Peterlongo, Alix Regnier‌​‌, Leo Ackermann,​​ Sebastien Bellenous, Khac​​​‌ Minh Tam Truong.‌

A challenge is to‌​‌ develop novel data representations​​ for genomic data across​​​‌ the spectrum of used‌ sequencing technologies. This involves‌​‌ implementing a controlled and​​ measured level of lossiness,​​​‌ ranging from lossless capture‌ to very sparse sketches.‌​‌

Another challenge is to​​ develop scalable methods for​​​‌ high-efficiency compression of genomic‌ data representations. The key‌​‌ idea is to leverage​​ the specific structural properties​​​‌ of genomic data, taking‌ their evolution over time‌​‌ into account. Following the​​ recent development of sequencing​​​‌ technologies, especially the advances‌ in metagenome-assembled genomes, we‌​‌ anticipate that in the​​ next ten years, we​​​‌ will first observe a‌ rapid increase of the‌​‌ total sequenced diversity, followed​​ by a saturation. We​​​‌ aim to develop computationally‌ tractable algorithms to quantify‌​‌ the geometric and repetitiveness​​ measures on public repositories​​​‌ such as SRA in‌ function of time, in‌​‌ order to parameterize the​​ data and exploit their​​​‌ redundancies. For instance, we‌ aim to leverage the‌​‌ phylogenetic relationships between individual​​ constituents of sequencing experiments​​​‌ (genomic, metagenomic, pangenomic, etc.)‌ to guide compression by‌​‌ our technique called phylogenetic​​ compression 1.

3.1.2​​​‌ Indexation

Participants: Karel Brinda‌, Pierre Peterlongo,‌​‌ Alix Regnier, Victor​​ Levallois, Sebastien Bellenous​​​‌.

The knowledge offered‌ by the sequencing data‌​‌ is priceless. Fortunately, a​​ large majority of the​​​‌ sequenced data is freely‌ accessible through large databanks‌​‌ such as ENA and​​ SRA, as previously mentioned.​​​‌ However, the raw sequences‌ stored in these data‌​‌ banks are not indexed​​ and therefore cannot be​​​‌ queried efficiently, apart from‌ direct accession lookups. With‌​‌ a few exceptions, these​​ data sets are never​​​‌ revisited.

Despite some recent‌ research efforts for indexing‌​‌ the raw sequencing data,​​ the most visible one​​​‌ being from Zurich group,‌ Switzerland 79, there‌​‌ exist currently no tool​​ able to provide a​​​‌ search engine able to‌ index hundreds of petabytes‌​‌ of data. State-of-the art​​ in this domain was​​​‌ proposed by the historical‌ GenScale team with kmtricks‌​‌ 5 and kmindex 4​​, used to index​​​‌ dozens of terabytes of‌ data and offering instant‌​‌ queries 2.

In​​ this context, our overall​​​‌ objective is to enable‌ the full and global‌​‌ exploitation of the generated​​ sequences, at the petabyte​​​‌ scale and growing exponentially,‌ allowing users to directly‌​‌ perform queries in raw​​ sequencing data on the​​​‌ fly in order to‌ tap into the largest‌​‌ underexploited resource in life​​ sciences. This is both​​​‌ a medium-term project, for‌ indexing the entire SRA‌​‌ database and also a​​​‌ long-term initiative to adapt​ to future changes in​‌ terms of both size​​ and required functionalities.

3.2.1 Complex​‌ variants & pangenome graph​​ analyses

Participants: Lune Angevin​​​‌, Siegfried Dubois,​ Claire Lemaitre, Fabrice​‌ Legeai, Melody Temperville​​, Riccardo Vicedomini.​​​‌

Phenotype variations, such as​ species adaptations, resistance to​‌ stress in plants, or​​ susceptibility to disease, are​​​‌ directly linked to genomic​ variations. Therefore, genomic variants​‌ have received considerable research​​ efforts for their detection​​​‌ and association with phenotypes.​ Most of the previous​‌ studies focused on small​​ variants (Single Nucleotide Polymorphisms​​​‌ or small insertions and​ deletions). Recent advances in​‌ sequencing technologies have highlighted​​ the prevalence of more​​​‌ complex variations, called structural​ variants (SVs) that are​‌ duplicated, deleted, inverted, or​​ displaced DNA segments, which​​​‌ cover 5 to 10​ times more bases than​‌ punctual ones 90.​​ Structural variants have been​​​‌ shown to have a​ major impact on various​‌ phenotypes 104, 63​​ and on the adaptation​​​‌ and evolution of species​ 105.

The rise​‌ of 3rd-generation sequencing (long​​ reads) has made it​​​‌ possible to characterize and​ catalog the full range​‌ of SVs in many​​ model organisms, such as​​​‌ in humans. However, their​ genome-wide detection remains difficult​‌ when it comes to​​ more complex genomes, such​​​‌ as polyploid ones, more​ divergent populations, and/or particular​‌ types of structural variants​​ such as those with​​​‌ highly repetitive features. Also,​ long-read technologies are still​‌ expensive and inaccessible for​​ re-sequencing projects where a​​​‌ large number of individuals​ are required. In these​‌ cases, short-read technologies are​​ still used either for​​​‌ the quantification problem (genotyping)​ or in association with​‌ long-distance information. We develop​​ methods for improving the​​​‌ detection and quantification of​ structural variants in such​‌ specific contexts and/or data​​ types.

A second area​​​‌ of research concerns genome​ representation models. The bioinformatics​‌ community is currently questioning​​ the dogma of the​​​‌ “reference sequence” to represent​ the genome of a​‌ species, given the large​​ number and diversity of​​​‌ variants detected within a​ single species 87.​‌ The term 'pangenome​​' is used to​​​‌ represent all the genomes​ and variants that characterize​‌ a species. The ideal​​ data structure for this​​​‌ representation is a sequence​ graph where each node​‌ represents part of one​​ or several genome(s), and​​​‌ each genome of the​ species can be read​‌ by traversing a particular​​ path in this graph.​​​‌ Several models and tools​ have already been proposed​‌ by the community 77​​, 73, but​​ many open problems remain.​​​‌ The construction of these‌ graphs currently relies on‌​‌ heuristics, as does their​​ analysis to characterize genetic​​​‌ variants within a population,‌ with results that vary‌​‌ widely from one method​​ to another. Furthermore, the​​​‌ transposition of classic genomic‌ analysis methods using the‌​‌ linear sequence of the​​ reference assembly to approaches​​​‌ that instead use a‌ reference pangenome in the‌​‌ form of a graph​​ of variations is not​​​‌ immediate. Finally, we work‌ on how to represent,‌​‌ detect and genotype structural​​ variants in such a​​​‌ graph, and in particular‌ when they are close,‌​‌ nested or mixed up​​ with the millions of​​​‌ other small punctual variants‌ 6.

3.2.2 Metagenome‌​‌ analyses

Participants: Roumen Andonov​​, Lune Angevin,​​​‌ Claire Lemaitre, Nicolas‌ Maurice, Pierre Peterlongo‌​‌, Emeline Roux,​​ Riccardo Vicedomini.

Metagenomics​​​‌ focuses on the analysis‌ of sequencing data derived‌​‌ from a mixture of​​ microorganisms characterizing an environment​​​‌ of interest 75.‌ The first step in‌​‌ understanding a microbial environment​​ is characterization of the​​​‌ organisms that are there.‌ This problem requires reconstructing‌​‌ the genomes of the​​ sequenced species (the metagenome​​​‌ assembly). We work‌ on this problem at‌​‌ different levels of granularity,​​ and we study the​​​‌ functional aspects of a‌ microbiome (e.g.,‌​‌ drug resistance, virulence, pathogenicity).​​

During metagenome assemblies, strains​​​‌ of the same species‌ (usually sharing high sequence‌​‌ identity) are often “collapsed”​​ into species-level consensus sequences,​​​‌ thus hiding strain variability‌ 102. Our objective‌​‌ is to provide a​​ strain-aware metagenome assembly brings​​​‌ along additional challenges: (i)‌ the actual number of‌​‌ strains is unknown; (ii)​​ strains of the same​​​‌ species might be highly‌ similar; (iii) the abundance‌​‌ can be so low​​ that it becomes hard​​​‌ to distinguish between rare‌ events and sequencing errors.‌​‌

We aim to develop​​ novel scalable methods for​​​‌ metagenome assembly (both at‌ species and strain resolution)‌​‌ in the case of​​ high-complexity microbial communities. We​​​‌ develop solutions able to‌ exploit (and possibly combine)‌​‌ the full potential of​​ different sequencing technologies such​​​‌ as short-read, long-read, but‌ also chromosome conformation capture-based‌​‌ technologies (e.g., Hi-C) in​​ order to reconstruct complex​​​‌ metagenomes more accurately at‌ both species and strain‌​‌ level. We also provide​​ formal characterizations of the​​​‌ strain resolution problem based,‌ for instance, on graph‌​‌ theory and combinatorial optimization​​ formulations.

3.2.3 Recover DNA​​​‌ storage data

Participants: Olivier‌ Boulle, Florestan de‌​‌ Moor, Dominique Lavenier​​, Julien Leblanc.​​​‌

The process of recovering‌ digital data stored on‌​‌ DNA involves a number​​ of steps, including a​​​‌ sequencing stage which, as‌ with genomics studies, generates‌​‌ a huge volume of​​ data that needs to​​​‌ be processed for downstream‌ analysis 68, 94‌​‌, 67.

The​​ aim of this axis​​​‌ of research is to‌ propose methods and algorithms‌​‌ that can process the​​ information output while sequencing​​​‌ DNA storing digital data‌ in a time compatible‌​‌ with the overall decoding​​ chain of DNA storage.​​​‌ Unlike biological data, the‌ structure of the information‌​‌ is known. This means​​​‌ that the usual bioinformatics​ algorithms can be highly​‌ optimized to increase their​​ efficiency. We work on​​​‌ this topic in collaboration​ with researchers who are​‌ proposing coding schemes specific​​ to DNA, so that​​​‌ the coding specificities can​ be used to accelerate​‌ the processing of sequencing​​ data. The overall objective​​​‌ is to enable decoding​ to be carried out​‌ on the fly.

3.3.1 Reduce data movements​

Participants: Charly Airault,​‌ Karel Brinda, Dominique​​ Lavenier, Meven Mognol​​​‌, Pierre Peterlongo.​

In digital systems, a​‌ large part of the​​ power consumption comes from​​​‌ the data movements between​ the computing units (CPU​‌ cores) and the main​​ memory (DIMM modules) 91​​​‌. Computing is cheap​ but transferring data to​‌ the computation unit is​​ expensive: reading a 32-bit​​​‌ word from the DRAM,​ for example, is 10,000​‌ times more energy-intensive than​​ a 32-bit integer addition!​​​‌ And genomic processing often​ has to navigate randomly​‌ through large data structures,​​ leading to low CPU​​​‌ cache efficiency.

One way​ of mitigating this effect​‌ is to design genomic​​ data structures that optimize​​​‌ the locality of data​ access. A typical example​‌ is the Bloom filter,​​ a data structure widely​​​‌ used in genomics to​ query the membership of​‌ an element in a​​ set. Because of their​​​‌ large size, they are​ generally not cacheable. Knowing​‌ whether an element is​​ present requires several accesses​​​‌ (a half-dozen) to an​ array, which translates into​‌ as many cache misses.​​ Optimized Bloom filters, at​​​‌ the cost of a​ slight degradation in precision,​‌ allow an element to​​ target the same memory​​​‌ zone, making much better​ use of caches and​‌ thus limiting data transfer​​ to the CPU 89​​​‌.

A second approach​ is the concept of​‌ “Processing-in-Memory” (PiM), where the​​ computing units are directly​​​‌ integrated into the main​ memory 72. In​‌ this case, data no​​ longer leave the memory,​​​‌ but are processed on​ the spot. In this​‌ way, numerous data movements​​ are avoided. The best​​​‌ example of this is​ probably querying a database​‌ stored in a PiM​​ memory: queries are broadcast​​​‌ to all the memory​ computing units, which process​‌ them simultaneously, and send​​ back only relevant data​​​‌ to the CPU 81​, 3. The​‌ PiM concept strongly limits​​ the exchanges between the​​​‌ processor cores and the​ main memory, and contributes​‌ to save energy.

To​​ sum-up, this axis has​​​‌ a long term objective​ of investigating how genomic​‌ data structures can be​​ designed to support locality​​​‌ of data access, and​ exploiting specialized hardware.

3.3.2​‌ Increase scaling, decrease computations​​

Participants: Pierre Peterlongo,​​​‌ Victor Levallois.

Our​ research is driven by​‌ biological questions involving the​​ analysis of large datasets.​​​‌ The tools we develop,​ independently from their optimizations,​‌ are designed to analyze​​ complete datasets and deliver​​​‌ extensive results. Yet there​ are many questions that​‌ can be answered without​​ the need for complete​​ data analysis, or exhaustive​​​‌ exploration of the associated‌ data structures.

In this‌​‌ context, various methods have​​ already proved successful. This​​​‌ is the case with‌ the sub-sampling of metagenomic‌​‌ data (SimkaMin 65,​​ Minhash 93) for​​​‌ comparing samples. This is‌ also applied for selecting‌​‌ some specific elements (such​​ as in PebbleScout 97​​​‌) from the full‌ datasets when the question‌​‌ is to index and​​ represent samples.

We exploit​​​‌ this concept following various‌ key ideas. 1/ to‌​‌ which extend can we​​ focus only on some​​​‌ elements (sub-sequences, specific portions‌ of graph data-structures, $\cdots ‌‌$) for answering a​​ biological question? 2/ conversely,​​​‌ determine and exclude large‌ sets of elements that‌​‌ do not carry information​​ for the question asked​​​‌ (ubiquitous sub-sequences, non variables‌ paths on graphs, $\cdots ‌‌$).

Insect genomics​​​‌ to reduce phytosanitary product‌ usage.

Through its long‌​‌ term collaboration with INRAE​​ IGEPP, GenScale is involved​​​‌ in various genomic projects‌ in the field of‌​‌ agricultural research. In particular,​​ we participate in the​​​‌ genome assembly and analyses‌ of some major agricultural‌​‌ pests or their natural​​ enemies such as parasitoids.​​​‌ The long term objective‌ of these genomic studies‌​‌ is to develop control​​ strategies to limit population​​​‌ outbreaks and the dissemination‌ of plant viruses they‌​‌ frequently transmit, while reducing​​ the use of phytosanitary​​​‌ products.

Energy efficient genomic‌ computation through Processing-in-Memory.

All‌​‌ current computing platforms are​​ designed following the von​​​‌ Neumann architecture principles, originated‌ in the 1940s, that‌​‌ separate computing units (CPU)​​ from memory and storage.​​​‌ Processing-in-memory (PIM) is expected‌ to fundamentally change the‌​‌ way we design computers​​ in the near future.​​​‌ These technologies consist of‌ processing capability tightly coupled‌​‌ with memory and storage​​ devices. As opposed to​​​‌ bringing all data into‌ a centralized processor, which‌​‌ is far away from​​ the data storage and​​​‌ is bottlenecked by the‌ latency (time to access),‌​‌ the bandwidth (data transfer​​​‌ throughput) to access this​ storage, and energy required​‌ to both transfer and​​ process the data, in-memory​​​‌ computing technologies enable processing​ of the data directly​‌ where it resides, without​​ requiring movement of the​​​‌ data, thereby greatly improving​ the performance and energy​‌ efficiency of processing of​​ massive amounts of data​​​‌ potentially by orders of​ magnitude. This technology is​‌ currently under test in​​ GenScale with a revolutionary​​​‌ memory component developed by​ the UPMEM company. Several​‌ genomic algorithms have been​​ parallelized on UPMEM systems,​​​‌ and we demonstrated significant​ energy gains compared to​‌ FPGA or GPU accelerators.​​ For comparable performances (in​​​‌ terms of execution time)​ on large scale genomics​‌ applications, UPMEM PIM systems​​ consume 3 to 5​​​‌ times less energy.

7.1.1 logan-search​

• Keywords:
  SRA, Indexing, Genomic​‌ sequence
• Functional Description:
  Given​​ a DNA sequence, the​​​‌ service replies in a​ few minutes in which​‌ SRA accession(s) it is​​ likely to occur. The​​​‌ service also enables to​ recover metadata associated to​‌ target accessions, and to​​ perform some visualizations. In​​​‌ more technical depth, the​ search engine uses kmindex,​‌ a k-mer based sequence​​ search tool that uses​​​‌ Bloom filters. It was​ applied to construct an​‌ index over all genome​​ assemblies of all of​​​‌ SRA, more specifically over​ the unitigs of Logan.​‌ This website is running​​ the kmviz visualization tool.​​​‌
• URL:
  https://­logan-search.­org/
• Publication:
  hal-05446815​
• Contact:
  Pierre Peterlongo
• Partners:​‌
  CEA, CHU Pasteur, University​​ of Toronto

7.1.2 muset​​

• Keywords:
  Unitig matrices, De​​​‌ Bruijn graphs, K-mer
• Functional‌ Description:
  MUSET is a‌​‌ software tool designed to​​ generate an abundance unitig​​​‌ matrix from a collection‌ of input samples in‌​‌ FASTA/Q format. It also​​ offers a comprehensive suite​​​‌ of tools for manipulating‌ k-mer matrices, along with‌​‌ a script for efficiently​​ generating a presence-absence unitig​​​‌ matrix.
• URL:
  https://­github.­com/­camiladuitama/­muset
• Contact:‌
  Riccardo Vicedomini
• Participants:
  Riccardo‌​‌ Vicedomini, Francesco Andreace, Yoann​​ Dufresne, Rayan Chikhi, Camila​​​‌ Duitama
• Partner:
  Sequence Bioinformatics‌

7.1.3 KmerCamel

• Name:
  KmerCamel‌​‌
• Keywords:
  Bioinformatics, Compression
• Functional​​ Description:
  KmerCamel provides implementations​​​‌ of several algorithms for‌ efficiently representing a set‌​‌ of k-mers as a​​ masked superstring.
• URL:
  https://­github.­com/­OndrejSladky/­kmercamel​​​‌
• Contact:
  Karel Brinda

7.1.4‌ Phylign

• Name:
  Phylign
• Keywords:‌​‌
  Bioinformatics, Alignment, Genomic sequence,​​ Data compression
• Functional Description:​​​‌
  A tool for rapid‌ BLAST-like search among 661k‌​‌ sequenced bacteria on personal​​ computers.
• URL:
  http://­github.­com/­karel-brinda/­mof-search
• Contact:​​​‌
  Karel Brinda
• Participant:
  Karel‌ Brinda
• Partners:
  European Bioinformatics‌​‌ Institute, HARVARD Medical School​​

7.1.5 Phylign-Fulgor

• Name:
  Phylign-Fulgor​​​‌
• Keywords:
  Bioinformatics, Genomics, Alignment,‌ Bacterial strains
• Functional Description:‌​‌
  Phylign-Fulgor is a Snakemake-based​​ pipeline that enables large-scale​​​‌ local alignment of genomic‌ queries against massive bacterial‌​‌ genome collections such as​​ the 661k and AllTheBacteria​​​‌ datasets on standard laptops‌ and desktops by replacing‌​‌ COBS with a customized​​ version of Fulgor and​​​‌ leveraging phylogenetic compression, which‌ reorganizes genomes according to‌​‌ evolutionary relationships to achieve​​ high compressibility, small index​​​‌ sizes, and efficient k-mer‌ search, it combines fast‌​‌ candidate selection using Fulgor​​ with accurate alignment using​​​‌ Minimap2, supports FASTA/FASTQ inputs,‌ scales through easy parallelization‌​‌ and optional cluster execution,​​ and provides a complete​​​‌ workflow from database download‌ and query preprocessing to‌​‌ matching, candidate aggregation, and​​ final alignments, allowing rapid,​​​‌ memory-efficient, and practical genomic‌ search across hundreds of‌​‌ thousands to millions of​​ bacterial genomes within hours.​​​‌
• URL:
  https://­github.­com/­Francii-B/­Phylign-Fulgor
• Contact:
  Karel‌ Brinda
• Participants:
  Karel Brinda,‌​‌ Francesca Brunetti

7.1.6 MiniPhy​​

• Name:
  MiniPhy
• Keywords:
  Compression,​​​‌ Bioinformatics, Genomic sequence, Data‌ compression
• Functional Description:
  Phylogenetic‌​‌ compression of extremely large​​ genome collections
• URL:
  https://­github.­com/­karel-brinda/­miniphy​​​‌
• Contact:
  Karel Brinda

7.1.7‌ FMSI

• Name:
  FMSI
• Keywords:‌​‌
  Bioinformatics, Genomics, Kmers, Indexing​​
• Functional Description:

  FMSI is​​​‌ a highly memory-efficient tool‌ for performing membership queries‌​‌ on single k -mer​​ sets. FMSI uses masked​​​‌ superstrings for storing the‌ k -mer sets to‌​‌ ensure high compressibility for​​ a wide range of​​​‌ different k -mer sets,‌ and implements FMS-index, a‌​‌ simplification of the FM-index.​​ It supports both streaming​​​‌ and single queries. The‌ functionality implemented in FMSI‌​‌ is based on the​​ following papers:

  The memory​​​‌ consumption for FMSI are‌ (w/o kLCP which is‌​‌ additional 1bit/superstring char):

  Queries:​​ 2.41 bits / canonical​​​‌ 31-mer with human genome,‌ bits / canonical 31-mer‌​‌ for E. coli pangenome​​ (1.17G k-mers from 89k​​​‌ genome), bits / canonical‌ 31-mer for SARS-CoV-2 Construction:‌​‌ 46 GB for human​​ genome

• Release Contributions:
  FMSI​​​‌ is able to perform‌ dictionary queries, i.e. to‌​‌ return for present k-mers​​ a hash value between​​​‌ 0 and |K|-1 (subcommand‌ hash) FMSI has simplified‌​‌ CLI The parameter for​​ path is now positional​​​‌ Removed some paramaters which‌ were not necessary Query‌​‌ streaming is by default​​​‌ disabled and can be​ turned on The output​‌ of FMSI is now​​ more detailed giving information​​​‌ not only about the​ percentage of found k-mers,​‌ but printing a presence​​ bitmap for each k-mer​​​‌ Due to the changes​ in CLI and output,​‌ this introduces breaking changes​​ compared to v0.3.x
• Contact:​​​‌
  Karel Brinda
• Participant:
  Karel​ Brinda
• Partner:
  Charles University​‌ Prague

7.1.8 strainberry2

• Keywords:​​
  Metagenome assembly, Long reads,​​​‌ Haplotype phasing
• Functional Description:​
  Strain-level metagenome assembly for​‌ long reads obtained with​​ modern sequencing technologies such​​​‌ as PacBio HiFi or​ Oxford Nanopore R10.
• News​‌ of the Year:
  Première​​ version du logiciel.
• URL:​​​‌
  https://­github.­com/­rvicedomini/­strainberry2
• Contact:
  Riccardo Vicedomini​

7.1.9 rs-pancat-compare

• Keyword:
  Pangenomics​‌
• Functional Description:
  Program that​​ calculates the distance between​​​‌ two GFA (Graphical Fragment​ Assembly) files. It takes​‌ in the file paths​​ of the two GFA​​​‌ files. The program first​ identifies the common paths​‌ between the two graphs​​ by finding the intersection​​​‌ of their path names.​ For each common path,​‌ the program reads those​​ and output differences in​​​‌ segmentation in-between them. The​ purpose is to output​‌ the necessary operations (merges​​ and splits) required to​​​‌ transform the graph represented​ by the first GFA​‌ file into the graph​​ represented by the second​​​‌ GFA file.
• URL:
  https://­github.­com/­dubssieg/­rs-pancat-compare​
• Publication:
  hal-04871087
• Contact:
  Siegfried​‌ Dubois
• Participants:
  Siegfried Dubois,​​ Claire Lemaitre
• Partner:
  INRAE​​​‌

7.1.10 Mapler

• Name:
  Metagenome​ Assembly and Evaluation Pipeline​‌ for Long Reads
• Keywords:​​
  Metagenomics, Genome assembly, Benchmarking,​​​‌ Bioinformatics
• Functional Description:
  Mapler​ is a pipeline to​‌ compare the performances of​​ long-read metagenomic assemblers. The​​​‌ pipeline is focused on​ assemblers for high fidelity​‌ long read sequencing data​​ (e.g. pacBio HiFi), but​​​‌ it supports also assemblers​ for low-fidelity long reads​‌ (ONT, PacBio CLR) and​​ hybrid assemblers. It currently​​​‌ compares metaMDBG, metaflye, Hifiasm-meta,​ opera-ms and miniasm as​‌ assembly tools, and uses​​ reference-based, reference-free and binning-based​​​‌ evalutation metrics. It is​ implemented in Snakemake.
• URL:​‌
  https://­gitlab.­inria.­fr/­mistic/­mapler
• Publication:
  hal-04142837
• Contact:​​
  Nicolas Maurice
• Participants:
  Nicolas​​​‌ Maurice, Claire Lemaitre, Riccardo​ Vicedomini, Clemence Frioux

7.1.11​‌ Alice

• Keywords:
  Genome assembly,​​ Haplotyping, NGS
• Functional Description:​​​‌
  Assemble DNA sequencing high-fidelity​ reads into full genomes.​‌ Based on the newly​​ introduced MSR sketching
• URL:​​​‌
  https://­github.­com/­rolandfaure/­alice-asm
• Contact:
  Roland Faure​

7.1.12 DnarXiv

• Name:
  dnarXiv​‌ project platform
• Keywords:
  Biological​​ sequences, Simulator, Sequence alignment,​​​‌ Error Correction Code
• Functional​ Description:
  The objective of​‌ DnarXiv is to implement​​ a complete system for​​​‌ storing, preserving and retrieving​ any type of digital​‌ document in DNA molecules.​​ The modules include the​​​‌ conversion of the document​ into DNA sequences, the​‌ use of error-correcting codes,​​ the simulation of the​​​‌ synthesis and assembly of​ DNA fragments, the simulation​‌ of the sequencing and​​ basecalling of DNA molecules,​​​‌ and the overall supervision​ of the system.
• URL:​‌
  https://­gitlab.­inria.­fr/­dnarxiv
• Contact:
  Olivier Boulle​​
• Participants:
  Olivier Boulle, Dominique​​​‌ Lavenier
• Partners:
  IMT Atlantique,​ Université de Rennes 1​‌

7.1.13 INVPG_annot

• Keywords:
  Pangenomics,​​ Structural Variation, Genomic sequence,​​​‌ Variation graphs
• Functional Description:​
  INVPG_annot is an automated​‌ method to identify, among​​ the bubbles found in​​​‌ a pangenome graph, the​ ones that correspond to​‌ specific inversion topologies. The​​ method starts with bubbles​​ that have already been​​​‌ found and aims at‌ annotating them, by analyzing‌​‌ their allele size, paths​​ and aligning allele sequences.​​​‌ INVPG-annot takes as input‌ a bubble file in‌​‌ VCF format, and outputs​​ a subset of these​​​‌ bubbles annotated as inversions‌ in VCF format.
• URL:‌​‌
  https://­github.­com/­SandraLouise/­INVPG_annot
• Publication:
  hal-05204329
• Contact:​​
  Claire Lemaitre
• Participants:
  Claire​​​‌ Lemaitre, Sandra Romain, Siegfried‌ Dubois

7.1.14 SVJedi-Tag

• Keywords:‌​‌
  Structural Variation, Variation graphs,​​ Genotyping
• Functional Description:
  SVJedi-Tag​​​‌ is a tool for‌ genotyping inversions using linked-read‌​‌ data. It is based​​ on the analysis of​​​‌ the distribution of barcode‌ signals on either sides‌​‌ of inversion breakpoints, with​​ inversions being represented in​​​‌ a variation graph. The‌ variation graph is built‌​‌ from a reference genome​​ and a VCF file​​​‌ containing the inversions to‌ be genotyped. VG giraffe‌​‌ is then used to​​ map the sample reads​​​‌ onto the graph. Then,‌ for each inversion, SVJedi-Tag‌​‌ analyzes the barcode signals​​ of reads aligned on​​​‌ each side of the‌ inversion breakpoints to estimate‌​‌ its allelic ratio and​​ predict its genotype.
• URL:​​​‌
  https://­github.­com/­Mtemperville/­SVJedi-Tag
• Publication:
  hal-05393269
• Contact:‌
  Melody Temperville
• Participants:
  Fabrice‌​‌ Legeai, Melody Temperville, Claire​​ Lemaitre

7.1.15 ConCluD

• Name:​​​‌
  ConCluD
• Keywords:
  Bioinformatics, Clustering,‌ Short reads, Long reads,‌​‌ DNA sequencing
• Functional Description:​​
  DNA-based data storage offers​​​‌ a compelling solution for‌ long-term, high-density archiving. In‌​‌ this framework, accurately recon­structing​​ high-quality encoded sequences after​​​‌ sequencing is critical, as‌ it directly impacts the‌​‌ design of error-correcting codes​​ optimized for DNA storage.​​​‌ Furthermore, efficient and scalable‌ processing is essential to‌​‌ manage the large volumes​​ of data expected in​​​‌ such applications. We introduce‌ a novel method based‌​‌ on de-Bruijn graph partitioning,​​ enabling fast and accurate​​​‌ processing of sequencing data‌ regard­ less of the‌​‌ underlying sequencing technology and​​ without requiring prior knowledge​​​‌ of the information encoded‌ in the oligonucleotides. It‌​‌ is implemented in C++​​ within the software ConCluD​​​‌ and optimized for multi-core‌ servers. It also provides‌​‌ Processing-in-Memory acceleration for the​​ UPMEM DIMMs hardware.
• URL:​​​‌
  https://­gitlab.­inria.­fr/­pim/­org.­pim.­dnarxiv
• Publication:
  hal-05375444
• Contact:‌
  Florestan De Moor
• Participants:‌​‌
  Dominique Lavenier, Florestan De​​ Moor, Olivier Boulle

7.1.16​​​‌ Kaminari

• Keywords:
  Genomics, Indexation,‌ Kmer, Algorithm, Search Engine,‌​‌ Computational biology, Biological sequences​​
• Functional Description:
  Kaminari is​​​‌ a genomic data indexing‌ tool. In other words,‌​‌ it allows you to​​ create a data structure​​​‌ from a set of‌ DNA documents (assembled genomes,‌​‌ sequencing reads, etc.). This​​ data structure can then​​​‌ be queried like a‌ search engine to find‌​‌ the origin of the​​ DNA sequence of interest.​​​‌
• Contact:
  Victor Levallois
• Participants:‌
  Victor Levallois, Pierre Peterlongo‌​‌

7.1.17 kmindex

• Keywords:
  Kmer,​​ Data structures, Indexing
• Functional​​​‌ Description:
  Given a databank‌ D = {S1, ...,‌​‌ Sn}, with each Si​​ being any genomic dataset​​​‌ (genome or raw reads),‌ kmindex allows to compute‌​‌ the percentage of shared​​ k-mers between a query​​​‌ Q and each S‌ in D. It supports‌​‌ multiple datasets and allows​​ searching for each sub-index​​​‌ Di in G =‌ {D1, ..., Dm}. Queries‌​‌ benefit from the findere​​ algorithm. In a few​​​‌ words, findere allows to‌ reduce the false positive‌​‌ rate at query time​​​‌ by querying (s+z)-mers instead​ of s-mers, which are​‌ the indexed words, usually​​ called k-mers. kmindex is​​​‌ a tool for querying​ sequencing samples indexed using​‌ kmtricks.
• News of the​​ Year:
  During the year​​​‌ 2025, the kmindex tool​ was vastly updated, mainly​‌ by Téo Lemane, including​​ the compression methods developed​​​‌ by Alix Regnier.
• URL:​
  https://­github.­com/­tlemane/­kmindex
• Contact:
  Pierre Peterlongo​‌
• Participants:
  Teo Lemane, Pierre​​ Peterlongo

7.1.18 back to​​​‌ sequences

• Keywords:
  Kmers, Genomic​ sequence
• Functional Description:
  “back​‌ to sequences” is a​​ software program that allows​​​‌ you to find the​ origin of words of​‌ size k (k-mers) in​​ raw sequencing data. This​​​‌ is a common operation​ when analyzing this type​‌ of data.
• News of​​ the Year:
  We updated​​​‌ in 2025 back to​ sequences. It is now​‌ faster, while accepting more​​ input file formats. It's​​​‌ deployment is now much​ easier.
• Contact:
  Pierre Peterlongo​‌

8.1.1 Phylogenetic​ compression

Participants: Karel Břinda​‌.

Comprehensive collections approaching​​ millions of sequenced genomes​​​‌ have become central information​ sources in the life​‌ sciences. However, the rapid​​ growth of these collections​​​‌ has made it effectively​ impossible to search these​‌ data using tools such​​ as the Basic Local​​​‌ Alignment Search Tool (BLAST)​ and its successors. Here,​‌ we present a technique​​ called phylogenetic compression, which​​​‌ uses evolutionary history to​ guide compression and efficiently​‌ search large collections of​​ microbial genomes using existing​​​‌ algorithms and data structures.​ We show that, when​‌ applied to modern diverse​​ collections approaching millions of​​​‌ genomes, lossless phylogenetic compression​ improves the compression ratios​‌ of assemblies, de Bruijn​​ graphs and k-mer indexes​​​‌ by one to two​ orders of magnitude (implemented​‌ in MiniPhy 7.1.6).​​ Additionally, we develop a​​​‌ pipeline 7.1.4 for a​ BLAST-like search over these​‌ phylogeny-compressed reference data, and​​ demonstrate it can align​​​‌ genes, plasmids or entire​ sequencing experiments against all​‌ sequenced bacteria until 2019​​ on ordinary desktop computers​​​‌ within a few hours.​ Phylogenetic compression has broad​‌ applications in computational biology​​ and may provide a​​​‌ fundamental design principle for​ future genomics infrastructure. 1​‌.

8.1.2 Logan Search,​​ a k-mer search engine​​​‌ for all Sequence Read​ Archive public accessions

Participants:​‌ Pierre Peterlongo.

There​​ are 50 petabases of​​​‌ freely-available DNA sequencing data.​ We proposed the yet​‌ unpublished Logan Search server​​ 46, built using​​​‌ and running thanks to​ kmtricks 85 and kmindex​‌ 83. This server​​ allows you to search​​​‌ for any DNA sequence​ in minutes, bringing Earth’s​‌ largest genomic resource to​​ your fingertips. Under the​​​‌ hood, we built a​ 1 petabyte k-mer index​‌ for all 27 million​​ sequencing datasets in the​​​‌ SRA up to 12-2023.​ Logan Search transforms any​‌ DNA query to its​​ $k$-mers (with $\mathrm{k‌}=31$), and​ it retrieves every dataset​‌ containing these $k$-mers.​​ It’s the only service​​​‌ working at this scale.​ The output datasets are​‌ visualized with custom plots​​ in Logan Search, which​​​‌ accesses a harmonized set​ of query and SRA​‌ meta-data including sequencing technology,​​ type of molecule, geographic​​ distribution, and sample origins.​​​‌ Fundamentally, Logan Search returns‌ a list of SRA‌​‌ accessions. To bring closer​​ to the data we​​​‌ also propose a microservice‌ to instantly retrieve contigs‌​‌ matching the search, also​​ visualisable thanks to a​​​‌ blast-like alignment.

8.1.3 Kaminari,‌ minimizing genomic index sizes‌​‌

Participants: Victor Levallois,​​ Pierre Peterlongo.

The​​​‌ problem of identifying the‌ set of textual documents‌​‌ from a given database​​ containing a query string​​​‌ has been studied in‌ various fields of computing,‌​‌ e.g., in Information Retrieval,​​ Databases, and Computational Biology.​​​‌ With Kaminari, we consider‌ the approximate version of‌​‌ this problem, that is,​​ the result set is​​​‌ allowed to contain some‌ false positive matches (but‌​‌ no false negatives), and​​ focus on the specific​​​‌ case where the indexed‌ documents are DNA strings.‌​‌ We explore an alternative​​ index design based on​​​‌ k-mers minimizers and integer‌ compression methods. We showed‌​‌ in 48 that a​​ careful implementation of this​​​‌ design outperforms previous solutions‌ based on Bloom filters‌​‌ by a wide margin:​​ the index has lower​​​‌ memory footprint and faster‌ query times, while false‌​‌ positive matches have only​​ a minor impact on​​​‌ the ranking of the‌ documents reported. This trend‌​‌ is robust across genomic​​ datasets of different complexity​​​‌ and query workloads.

8.1.4‌ K-mer set representation using‌​‌ masked superstrings

Participants: Karel​​ Břinda.

The popularity​​​‌ of k-mer-based methods has‌ recently driven the development‌​‌ of compact k-mer-set representations​​ such as simplitigs/Spectrum-Preserving String​​​‌ Sets (SPSS), matchtigs, and‌ eulertigs, which represent k-mer‌​‌ sets via strings that​​ contain individual k-mers as​​​‌ substrings more efficiently than‌ traditional unitigs. Previously, we‌​‌ showed that all these​​ representations correspond to superstrings​​​‌ of the input k-mers‌ and can be generalized‌​‌ within a unified framework​​ called masked superstrings of​​​‌ k-mers, analyzed the computational‌ complexity of this framework,‌​‌ prove NP-hardness for computing​​ both k-mer superstrings and​​​‌ their masks, and design‌ local and global greedy‌​‌ heuristics for practical computation​​ 99 . We implemented​​​‌ these heuristics in the‌ KmerCamel 7.1.3 tool and‌​‌ demonstrate that masked superstrings​​ unify the theory and​​​‌ practice of textual k-mer‌ set representations while providing‌​‌ a flexible basis for​​ application-specific optimization. In 2025,​​​‌ we focused primarily on‌ the optimization of the‌​‌ approximation algorithms, in particular​​ for greedy approximation for​​​‌ simplitigs/Spectrum-Preserving String Sets, which‌ resulted in new major‌​‌ releases of KmerCamel 7.1.3​​.

Building on this​​​‌ foundation, we addressed the‌ need for scalable and‌​‌ versatile indexes for arbitrary​​ k-mer sets in the​​​‌ face of rapidly growing‌ and heterogeneous genomic data‌​‌ vian an index called​​ FMSI. Unlike state-of-the-art indexes​​​‌ such as SBWT and‌ SSHash, whose performance relies‌​‌ on long non-branching paths​​ in de Bruijn graphs,​​​‌ FMSI adopts a superstring-based‌ design that supports efficient‌​‌ membership and compressed dictionary​​ queries with strong theoretical​​​‌ guarantees. FMSI exploits recent‌ advances in k-mer superstrings‌​‌ together with the Masked​​ Burrows–Wheeler Transform, which extends​​​‌ the classical Burrows–Wheeler Transform‌ by incorporating position masking.‌​‌ Across genomic, pangenomic, and​​ metagenomic datasets and a​​​‌ wide range of k‌ values, FMSI consistently improves‌​‌ query space efficiency, often​​​‌ using 2–3× less memory​ than competing methods while​‌ maintaining comparable or faster​​ query times. 19

Finally,​​​‌ we further extended the​ masked superstring framework to​‌ support dynamic operations on​​ large k-mer sets, a​​​‌ setting that static compact​ representations cannot handle efficiently.​‌ We introduced f-masked superstrings,​​ which combine masked superstrings​​​‌ with custom demasking functions​ to enable k-mer set​‌ operations through index merging.​​ By integrating this concept​​​‌ withi the FMSI index,​ we obtained a memory-efficient​‌ k-mer membership index and​​ compressed dictionary that support​​​‌ set operations via Burrows–Wheeler​ Transform merging. This approach​‌ offers a promising theoretical​​ solution to dynamic k-mer​​​‌ set manipulation and highlights​ f-masked superstrings as a​‌ strong candidate for an​​ elementary data type for​​​‌ k-mer sets.sladky: 24

The​ masked superstring framework was​‌ presented RECOMB-2025 59 and​​ DSB 2025 36.​​​‌

8.1.5 Optimized k-mer search​ across millions of bacterial​‌ genomes on laptops

Participants:​​ Karel Břinda, Francesca​​​‌ Brunetti.

Comprehensive bacterial​ collections have reached millions​‌ of genomes, opening new​​ opportunities for point-of-care diagnostics​​​‌ and epidemiological surveillance. However,​ local real-time search over​‌ such collections on commodity​​ hardware remains difficult. Currently,​​​‌ only LexicMap and Phylign​ enable local search and​‌ alignment at such a​​ scale; among them, only​​​‌ Phylign is designed to​ run on laptops, via​‌ a subindex approach informed​​ by phylogenetic compression. However,​​​‌ Phylign's performance deteriorates on​ long and divergent queries​‌ because it uses COBS​​ as a k-mer-based prefilter​​​‌ before alignment with Minimap2.​ Meanwhile, recent k -mer​‌ indexes such as Fulgor​​ and Themisto have emerged,​​​‌ but there is no​ practical methodology for selecting,​‌ combining, and parameterizing them​​ for phylogenetically partitioned million-genome​​​‌ search under constraints. Here,​ we develop an end-to-end​‌ methodology for k-mer matching​​ in phylogenetically compressed bacterial​​​‌ collections. We formalize a​ matching strategy defined by​‌ matching mode, query type,​​ and reference characteristics, and​​​‌ use this to shortlist​ candidate indexes and benchmark​‌ them under space–time trade-offs.​​ As a case study,​​​‌ we address plasmid search​ over AllTheBacteria, compare multiple​‌ index types, and identify​​ configurations optimizing the Pareto​​​‌ frontier of space and​ speed. Guided by these​‌ results, we implement a​​ phylogenetically compressed variant of​​​‌ Fulgor, integrate it into​ Phylign, and obtain Phylign-Fulgor,​‌ a laptop-ready pipeline for​​ million-genome search. On the​​​‌ 661k collection, Phylign-Fulgor makes​ the prefiltering step  4×​‌ faster than Phylign-COBS at​​ the cost of a​​​‌ 1.2× larger index. On​ AllTheBacteria, its k-mer filter​‌ is 20x–300x faster in​​ real time than LexicMap’s​​​‌ alignment-based search and uses​  20× smaller disk space.​‌ The full Phylign-Fulgor workflow​​ including Minimap2 alignments is​​​‌ slower than LexicMap for​ a single plasmid but​‌ competitive or faster for​​ batched plasmid queries. Phylign-Fulgor​​​‌ has comparable matching sensitivity​ to LexicMap, is less​‌ sensitive at the alignment​​ level, but always stays​​​‌ within a laptop RAM​ budget ( 5×–20× lower​‌ memory than LexicMap). 44​​ The work was presented​​​‌ at DSB 2025 31​ and 53rd National Congress​‌ of the Italian Society​​ of Microbiology 32.​​​‌

8.1.6 k-mer matrix compression​

Participants: Pierre Peterlongo,​‌ Alix Regnier.

In​​ the work 35 presented​​ at DSB2025, we​​​‌ propose a block compression‌ method on top of‌​‌ the kmtricks matrices 84​​, when used by​​​‌ kmindex for indexing and‌ query purposes. The main‌​‌ objective is to achieve​​ sensible compression ratios with​​​‌ a limited impact on‌ the index creation time‌​‌ and on the query​​ time. This method enables​​​‌ partial and targeted decompression,‌ thus optimizing data processing.‌​‌ To improve the compressibility​​ of matrices, we also​​​‌ exploit an idea inspired‌ by phylogenetic compression 8‌​‌. Hence, the impact​​ of reordering matrix columns​​​‌ based on phylogenetic order‌ is studied and used‌​‌ as a means of​​ increasing their compressibility. Preliminary​​​‌ results show that this‌ reordering significantly improves matrix‌​‌ compressibility. Finally, the approach​​ we propose reduces the​​​‌ storage space required, with‌ a limited impact on‌​‌ query time, making k-mer​​ matrices more accessible and​​​‌ usable.

8.1.7 Optimized phylogenetic‌ batching of million-genome collections‌​‌ for reduced storage requirements​​ and faster data retrieval​​​‌

Participants: Tam Truong,‌ Dominique Lavenier, Pierre‌​‌ Peterlongo, Karel Břinda​​.

Efficient compression is​​​‌ essential for the practical‌ use of million-genome bacterial‌​‌ collections, yet increasing scale​​ and diversity make this​​​‌ progressively harder. The most‌ effective approach to date‌​‌ is phylogenetic compression, which​​ orders genomes according to​​​‌ evolutionary relationships, as implemented‌ in MiniPhy. MiniPhy groups‌​‌ genomes into phylogenetically related​​ batches of limited size​​​‌ and then applies local‌ phylogenetic reordering prior to‌​‌ compression, forming the basis​​ of AllTheBacteria (n =​​​‌ 2,440,377) and achieving a‌ 40-fold reduction compared to‌​‌ gzip. However, MiniPhy relies​​ on heuristic batching based​​​‌ on accession order within‌ species, used as a‌​‌ proxy for similarity. While​​ sufficient for smaller datasets,​​​‌ this assumption breaks down‌ at million-genome scale, where‌​‌ individual species can exceed​​ hundreds of thousands of​​​‌ genomes and related genomes‌ are no longer close‌​‌ in accession order, leading​​ to reduced compression efficiency.​​​‌ In addition, MiniPhy is‌ tightly coupled to XZ,‌​‌ limiting the use of​​ newer compressors that offer​​​‌ comparable or better compression‌ with substantially faster performance.‌​‌ We introduce a new​​ phylogenetically informed batching strategy​​​‌ based on approximate species-level‌ phylogeny reconstruction, enabling effective‌​‌ co-location of related genomes​​ even at million-genome scale.​​​‌ Applied to AllTheBacteria excluding‌ rare or unknown species‌​‌ (n = 2,282k), this​​ yields batches over 40%​​​‌ more compressible with XZ‌ (20% when all genomes‌​‌ are included). Using MBGC​​ further reduces the compressed​​​‌ size to 15 GB‌ (41.3 GB for the‌​‌ full collection), and combining​​ this with optimized AGC​​​‌ results in an 18×‌ improvement in random-access single-genome‌​‌ retrieval time (0.5 s​​ versus 8.9–9.5 s). 38​​​‌

8.1.8 Towards space-efficient data‌ structures for large genome-distance‌​‌ matrices with quick retrieval​​

Participants: Léo Ackermann,​​​‌ Karel Břinda, Pierre‌ Peterlongo.

Many standard‌​‌ bioinformatics analyses lean on​​ pairwise distances between genomes.​​​‌ As a result, the‌ scalability of multiple downstream‌​‌ analyses relies on the​​ efficient computation and storage​​​‌ of pairwise distance matrices.‌ However, while the efficient‌​‌ computation of distances has​​ been addressed by modern​​​‌ sketching-based methods such as‌ Mash 93 and successors,‌​‌ the storage and indexing​​​‌ of the resulting matrices​ remain a significant challenge.​‌ In fact, due to​​ their quadratic size in​​​‌ the number of genomes,​ these matrices already surpass​‌ most storage capacities and​​ are thus heavily truncated​​​‌ when stored. This calls​ for a dedicated data​‌ structure that would be​​ space-efficient and support near-constant-time​​​‌ distance retrieval queries. In​ our work, we focused​‌ on showed that theoretical​​ collections of genomes model​​​‌ can be stored in​ linear space supporting constant-time​‌ queries. We then demonstrated​​ our preliminary results on​​​‌ the tradeoffs that exist​ between metric distortion and​‌ storing cost in practical​​ use cases. Preliminary results​​​‌ were presented at DSB2025​ 25.

8.2.1​​​‌ Pairwise graph edit distance​ characterizes the impact of​‌ the construction method on​​ pangenome graphs

Participants: Siegfried​​​‌ Dubois, Claire Lemaitre​.

Pangenome variation graphs​‌ are an increasingly used​​ tool to perform genome​​​‌ analysis, aiming to replace​ a linear reference in​‌ a wide variety of​​ genomic analyses. The construction​​​‌ of a variation graph​ from a collection of​‌ chromosome-size genome sequences is​​ a difficult task that​​​‌ is generally addressed using​ a number of heuristics.​‌ The question that arises​​ is to what extent​​​‌ the construction method influences​ the resulting graph, and​‌ the characterization of variability.​​ We aimed at characterizing​​​‌ the differences between variation​ graphs derived from the​‌ same set of genomes​​ with a metric which​​​‌ expresses and pinpoint differences.​ We designed a pairwise​‌ variation graph comparison algorithm,​​ which establishes an edit​​​‌ distance between variation graphs,​ threading the genomes through​‌ both graphs. We applied​​ our method to pangenome​​​‌ graphs built from yeast​ and human chromosome collections,​‌ and demonstrated that our​​ method effectively characterizes discordances​​​‌ between pangenome graph construction​ methods and scales to​‌ real datasets 11.​​

pancat compare is published​​​‌ as free Rust software​ under the AGPL3.0 open​‌ source license. Source code​​ and documentation are publicly​​​‌ available.

8.2.2 SVJedi-Tag: a​ novel method for genotyping​‌ large inversions with linked-read​​ data

Participants: Fabrice Legeai​​​‌, Claire Lemaitre,​ Mélody temperville.

Structural​‌ Variants (SVs) are an​​ important but overlooked aspect​​​‌ of genetic variation. In​ particular, inversions are known​‌ for their role in​​ the evolution of biological​​​‌ diversity and particularly studied​ in non-model species using​‌ population data. One of​​ the major steps in​​​‌ the study of SVs​ is genotyping. Linkedread data​‌ provide a cost-efficient alternative​​ to long-reads to genotype​​​‌ many individuals, by combining​ the low sequencing cost​‌ of short reads with​​ long-distance information thanks to​​​‌ the use of barcodes​ tagging long molecules. Whereas​‌ several methods have been​​ proposed to discover SVs​​​‌ with linked-reads, there are​ currently no tool for​‌ genotyping with this type​​ of sequencing data. In​​​‌ this paper, we present​ SVJedi-Tag, the first inversion​‌ genotyping method dedicated to​​ linked-read data. We tested​​​‌ SVJedi-Tag on simulated and​ real linked-read data in​‌ the seaweed fly Coelopa​​ frigida, and showed​​​‌ that SVJedi-Tag is able​ to genotype with high​‌ accuracy large inversions above​​ 25 kb, with a​​ read depth as low​​​‌ as 3X 37.‌

8.2.3 Mapler: a pipeline‌​‌ for assessing assembly quality​​ in taxonomically rich metagenomes​​​‌ sequenced with HiFi reads‌

Participants: Claire Lemaitre,‌​‌ Nicolas Maurice, Riccardo​​ Vicedomini.

Metagenome assembly​​​‌ seeks to reconstruct the‌ most high-quality genomes from‌​‌ sequencing data of microbial​​ ecosystems. Despite technological advancements​​​‌ that facilitate assembly, such‌ as Hi-Fi long reads,‌​‌ the process remains challenging​​ in complex environmental samples​​​‌ consisting of hundreds to‌ thousands of populations.

We‌​‌ designed and implemented Mapler​​ 14, 58,​​​‌ a metagenomic assembly and‌ evaluation pipeline with a‌​‌ primary focus on evaluating​​ the quality of HiFi-based​​​‌ metagenome assemblies. It integrates‌ several state-of-the-art tools as‌​‌ well as novel metrics​​ and visualizations based on​​​‌ read-to-contig alignments. It provides‌ a broad view of‌​‌ the sequence characteristics after​​ assembly and binning, in​​​‌ order to identify the‌ bottlenecks faced during bioinformatic‌​‌ processes. It is implemented​​ in an easy-to-use and​​​‌ customizable workflow. We applied‌ Mapler to three metagenomic‌​‌ datasets of increasing complexity,​​ all sequenced using PacBio​​​‌ HiFi technology: a mock‌ community of 21 populations,‌​‌ a human gut microbiome,​​ and a deadwood microbiome,​​​‌ demonstrating that it is‌ an effective way to‌​‌ examine assembly in taxonomically​​ rich ecosystems.

Mapler is​​​‌ open source and publicly‌ available.

8.2.4 Investigating pre-assembly‌​‌ clustering of PacBio HiFi​​ reads for de novo​​​‌ assembly of complex metagenomes‌

Participants: Claire Lemaitre,‌​‌ Nicolas Maurice, Riccardo​​ Vicedomini.

Metagenome assembly​​​‌ remains difficult in taxonomically‌ complex ecosystems, where low-coverage‌​‌ species and graph complexity​​ limit assembly quality. Although​​​‌ clustering reads prior to‌ assembly can reduce graph‌​‌ complexity, it may also​​ lower the effective coverage​​​‌ and has therefore seen‌ limited adoption. This work‌​‌ explored whether upstream clustering​​ of PacBio HiFi long​​​‌ reads can effectively improve‌ metagenomic assembly.

We compared‌​‌ the assembly of individual​​ simulated genomes against those​​​‌ same simulated genomes introduced‌ within a complex metagenome‌​‌ containing related species. We​​ confirmed that genomes are​​​‌ better assembled individually than‌ within the metagenome. Building‌​‌ on this observation, we​​ proposed an iterative clustering-based​​​‌ assembly framework. Reads were‌ first assembled within individual‌​‌ clusters. A complete assembly​​ was than generated through​​​‌ a bottom-up hierarchical approach‌ which aims at recovering‌​‌ mis-clustered or unassembled reads​​ (e.g., reintroducing​​​‌ them into neighboring clusters).‌ The method is modular‌​‌ and compatible with different​​ clustering strategies and assemblers,​​​‌ and can also operate‌ without iteration by simply‌​‌ merging cluster-wise assemblies.

We​​ evaluated this approach on​​​‌ a complex PacBio HiFi‌ soil metagenomic dataset 52‌​‌, 42 using both​​ coverage-based and taxonomy-driven clustering​​​‌ strategies. While clustering reduced‌ the fraction of reads‌​‌ aligning back to the​​ final assembly and the​​​‌ iterative rescue yielded only‌ marginal improvements, several genomes‌​‌ recovered through clustering-based strategies​​ showed higher quality in​​​‌ terms of completeness, contamination,‌ and contiguity. These results‌​‌ indicate that, despite current​​ limitations in pre-assembly clustering​​​‌ and read rescue, clustering-based‌ approaches could in principle‌​‌ enhance the recovery of​​ specific genomes and warrant​​​‌ further methodological improvements 33‌.

8.2.5 MUSET: set‌​‌ of utilities for constructing​​​‌ abundance unitig matrices from​ sequencing data

Participants: Riccardo​‌ Vicedomini.

Unitigs are​​ biological sequences that compactly​​​‌ and exhaustively represent sequencing​ data or assembled genomes.​‌ They are constructed from​​ k-mers but they avoid​​​‌ the redundancy of sequences​ covering the same genomic​‌ locus. A unitig matrix​​ is a data structure​​​‌ representing sequence content across​ multiple experiments by recording​‌ a numerical value for​​ each unitig across all​​​‌ samples. Unitig matrices extend​ the concept of k-mer​‌ matrices, which are gaining​​ popularity for sequence-phenotype association​​​‌ studies. They additionally preserve​ variations between samples while,​‌ at the same time,​​ significantly reducing disk space​​​‌ compared to k-mer matrices.​ We addressed the limitations​‌ of current methodologies by​​ developing MUSET 21,​​​‌ a method that integrates​ efficient k-mer counting and​‌ unitig extraction in order​​ to generate unitig matrices​​​‌ containing abundance values across​ the input samples. MUSET​‌ overcomes the limitation of​​ state-of-the-art methods which could​​​‌ only output binary (​i.e., presence-absence) matrices.​‌ We applied MUSET to​​ a 618-GB collection of​​​‌ ancient oral sequencing samples,​ for which it is​‌ able to build an​​ abundance-based unitig matrix in​​​‌ less than 10 hours​ and using 20 GB​‌ of memory. MUSET is​​ expected to facilitate the​​​‌ extraction of biologically significant​ sequences, making it a​‌ valuable contribution to downstream​​ sequencing data analyses such​​​‌ as genome-wide (or metagenome-wide)​ association studies.

MUSET is​‌ open source and publicly​​ available.

8.2.6 Improved strain-level​​​‌ metagenome assembly for modern​ long reads

Participants: Riccardo​‌ Vicedomini.

Recent long-read​​ sequencing technologies (PacBio HiFi​​​‌ and ONT R10.4) have​ improved metagenome assembly but​‌ still struggle to resolve​​ highly similar bacterial strains​​​‌ (sequence identity greater than​ 99%). We developed Strainberry2​‌ 39, an improved​​ strain-aware post-processing method that​​​‌ reconstructs strain-specific sequences from​ a species-level assembly and​‌ accurate long reads. Strainberry2​​ combines read dereplication with​​​‌ an iterative strain-phasing algorithm​ based on co-occurring single-nucleotide​‌ variants. Evaluated on 64​​ simulated datasets and on​​​‌ the ZymoBIOMICS Gut Microbiome​ Standard mock community, Strainberry2​‌ consistently reduced strain-level errors​​ (2-7X fewer misassemblies and​​​‌ at least 10X fewer​ strain-switch errors) compared to​‌ state-of-the-art assemblers. With PacBio​​ HiFi data, it matched​​​‌ the accuracy of hifiasm-meta​ while achieving a more​‌ balanced trade-off between contiguity​​ and duplication.

Strainberry2 is​​​‌ open source and publicly​ available.

8.2.7 Strain-level metagenomic​‌ classification through assembly-driven database​​ reduction

Participants: Josipa Lipovac​​​‌, Riccardo Vicedomini.​

Strain-level metagenomic classification is​‌ essential for understanding microbial​​ diversity and functional potential,​​​‌ but remains challenging, particularly​ in the absence of​‌ prior knowledge about the​​ composition of the sample.​​​‌ We developed MADRe 49​, a modular and​‌ scalable pipeline for long-read​​ strain-level metagenomic classification, enhanced​​​‌ with Metagenome Assembly-Driven Database​ Reduction. MADRe combines long-read​‌ metagenome assembly, contig-to-reference mapping​​ reassignment based on an​​​‌ expectation-maximization algorithm for database​ reduction, and probabilistic read​‌ mapping reassignment to achieve​​ sensitive and precise classification.​​​‌ We extensively evaluated MADRe​ on simulated datasets, mock​‌ communities, and a real​​ anaerobic digester sludge metagenome,​​​‌ demonstrating that it consistently​ outperforms existing tools by​‌ achieving higher precision with​​ reduced false positives. MADRe's​​ design allows users to​​​‌ apply either the database‌ reduction or read classification‌​‌ step individually. Using only​​ the read classification step​​​‌ shows results on par‌ with other tested tools.‌​‌

8.2.8 Sequencing data processing​​ for DNA storage

Participants:​​​‌ Florestan de Moor,‌ Olivier Boullé, Dominique‌​‌ Lavenier.

DNA-based data​​ storage presents an innovative​​​‌ and effective approach for‌ long-term, high-density archiving. Within‌​‌ this context, the precise​​ reconstruction of encoded sequences​​​‌ after sequencing is of‌ paramount importance, as it‌​‌ directly shapes the development​​ of error-correcting codes tailored​​​‌ for DNA storage. Additionally,‌ the ability to process‌​‌ data efficiently and on​​ a large scale is​​​‌ crucial to handle the‌ vast amounts of information‌​‌ anticipated in these applications.​​

We have developed a​​​‌ new method based on‌ de Bruijn graph partitioning.‌​‌ This approach enables rapid​​ and accurate processing of​​​‌ sequencing data, regardless of‌ the sequencing technology used,‌​‌ and does not require​​ prior knowledge of the​​​‌ information encoded in the‌ oligonucleotides. When tested on‌​‌ both synthetic and real-world​​ datasets, the method demonstrates​​​‌ outstanding precision and recall‌ 15

The method is‌​‌ implemented in C++ as​​ part of the ConCluD​​​‌ software, which is optimized‌ for multi-core servers. Our‌​‌ experiments reveal that a​​ dataset comprising 89 million​​​‌ reads, equivalent to a‌ 10 GB FASTA file,‌​‌ can be fully processed​​ in less than one​​​‌ minute on a standard‌ server with 32 cores.‌​‌

8.2.9 Data encoding for​​ DNA storage

Participants: Dominique​​​‌ Lavenier.

We have‌ proposed a novel encoding‌​‌ method for storing encrypted​​ data in DNA form,​​​‌ addressing key biological constraints‌ such as G-C content,‌​‌ homopolymers, and prohibited nucleotide​​ motifs used for data​​​‌ indexing. Our approach is‌ built on two innovations:‌​‌ first, a dynamic encoding​​ (DE) mechanism that utilizes​​​‌ variable-length DNA codewords to‌ encode binary data while‌​‌ avoiding homopolymers longer than​​ N bases; second, a​​​‌ sliding window strategy that‌ detects prohibited motifs in‌​‌ real time and inserts​​ non-coding bases to prevent​​​‌ their formation. Unlike existing‌ schemes, our method scales‌​‌ effortlessly for high homopolymer​​ thresholds and large sets​​​‌ of prohibited motifs, and‌ it remains independent of‌​‌ the cryptosystem used. We​​ provide a theoretical information​​​‌ rate for our proposal,‌ demonstrating significantly higher performance—particularly‌​‌ in terms of information​​ rate—compared to existing schemes,​​​‌ while maintaining the desired‌ G-C content with extremely‌​‌ high probability. Experimental validation​​ using a DNA storage​​​‌ chain simulator confirms that‌ our method requires a‌​‌ comparable number of sequence​​ copies to other encoding​​​‌ strategies for error-free data‌ recovery, underscoring its robustness‌​‌ and efficiency for secure,​​ long-term DNA-based data storage​​​‌ 7.

8.3.1 Processing-in-Memory: protein database​​ search

Participants: Charly Airault​​​‌, Florestan de Moor‌, Erwan Drezen,‌​‌ Meven Mognol, Dominique​​ Lavenier.

We investigated​​​‌ a massively parallel approach‌ for accelerating protein database‌​‌ search using a Processing-in-Memory​​ (PiM) architecture, leveraging UPMEM’s​​​‌ specialized DRAM technology. The‌ study focuses on parallelizing‌​‌ sequence alignment searches within​​ large protein databases, such​​​‌ as UniProt/SwissProt, by distributing‌ the database across PiM‌​‌ modules and broadcasting queries​​​‌ to each processing unit.​ The PiM-based implementation achieves​‌ significant speedups—up to 47.8x​​ faster than traditional multi-core​​​‌ servers—while maintaining result quality​ comparable to the widely​‌ used BLASTp tool. Energy​​ efficiency is also substantially​​​‌ improved, with an 8-core​ server equipped with 16​‌ PiM modules consuming 13​​ times less energy for​​​‌ equivalent workloads. Future work​ includes extending the method​‌ to support gapped alignments​​ and integrating advanced search​​​‌ heuristics. This research underscores​ the potential of PiM​‌ technology to revolutionize computationally​​ intensive genomic data processing​​​‌ 26.

8.3.2 Processing-in-Memory:​ energy efficiency

Participants: Florestan​‌ de Moor, Erwan​​ Drezen, Meven Mognol​​​‌, Dominique Lavenier.​

This work evaluates the​‌ energy efficiency of genomics​​ algorithms using Processing-in-Memory (PiM)​​​‌ architectures, which reduce data​ movement bottlenecks by integrating​‌ computation directly within memory.​​ Focusing on the UPMEM​​​‌ PiM platform, we assess​ the performance of key​‌ genomic tasks—sorting, data compression,​​ mapping, sequence comparison, and​​​‌ protein database search—on a​ server equipped with 2,560​‌ DPUs. Results show that​​ PiM delivers significant energy​​​‌ savings (up to 10.5x)​ and speedups (up to​‌ 15.2x) for applications with​​ fine-grained parallelism and irregular​​​‌ memory access patterns, such​ as protein database searches.​‌ However, tasks like sorting,​​ which benefit from optimized​​​‌ CPU cache and SIMD​ instructions, see limited gains.​‌ Our findings highlight PiM’s​​ potential to enhance scalability​​​‌ and sustainability in high-throughput​ bioinformatics, particularly for data-intensive​‌ workflows where traditional architectures​​ struggle with memory bandwidth​​​‌ limitations 34.

8.4.1 Investigating the topological​​​‌ motifs of inversions in​ pangenome graphs

Participants: Siegfried​‌ Dubois, Fabrice Legeai​​, Claire Lemaitre.​​​‌

Recent technological advances have​ accelerated the production of​‌ high-quality genome assemblies within​​ species, driving the growing​​​‌ use of pangenome graphs​ in genetic diversity analyses.​‌ These graphs reduce reference​​ bias in read mapping​​​‌ and enhance variant discovery​ and genotyping from SNPs​‌ to Structural Variants (SVs).​​ In pangenome graphs, variants​​​‌ appear as bubbles, which​ can be detected by​‌ dedicated bubble calling tools.​​ Although these tools report​​​‌ essential information on the​ variant bubbles, such as​‌ their position and allele​​ walks in the graph,​​​‌ they do not annotate​ the type of the​‌ detected variants. While simple​​ SNPs, insertions, and deletions​​​‌ are easily distinguishable by​ allele size, large balanced​‌ variants like inversions are​​ harder to differentiate among​​​‌ the large number of​ unannotated bubbles. In fact,​‌ inversions and other types​​ of large variants remain​​​‌ underexplored in pangenome graph​ benchmarks and analyses.

In​‌ this work we focused​​ on inversions, which have​​​‌ been drawing renewed attention​ in evolutionary genomics studies​‌ in the past years,​​ and aimed to assess​​​‌ how this type of​ variant is handled by​‌ state of the art​​ pangenome graphs pipelines. We​​​‌ identified two distinct topological​ motifs for inversion bubbles:​‌ one path-explicit and one​​ alignment-rescued. We developed INVPG_annot​​​‌ an automated method to​ identify, among the bubbles​‌ found in a pangenome​​ graph, the ones that​​ correspond to these specific​​​‌ inversion topologies. We constructed‌ pangenome graphs with both‌​‌ simulated data and real​​ data using four state​​​‌ of the art pipelines,‌ and assessed the impact‌​‌ of inversion size, punctual​​ genome divergence and haplotype​​​‌ number on inversion representation‌ and accuracy.

Our results‌​‌ reveal substantial differences between​​ pipelines, with many inversions​​​‌ either misrepresented or lost,‌ highlighting major challenges in‌​‌ analyzing inversions in pangenomic​​ approaches 50.

8.4.2​​​‌ Detecting chromosomal inversions for‌ population genomics: what could‌​‌ be the optimal approach?​​

Participants: Claire Lemaitre,​​​‌ Mélody Temperville.

Chromosomal‌ inversion has aroused a‌​‌ growing interest as it​​ plays a critical role​​​‌ in evolutionary processes. To‌ understand the genetic architecture‌​‌ of adaptation and the​​ importance of inversions, it​​​‌ is important to characterize‌ the full range of‌​‌ inversions accurately and in​​ a cost-efficient way across​​​‌ individuals and populations. We‌ addressed this objective by‌​‌ comparing the effectiveness of​​ characterizing inversions at population-scale​​​‌ using two sequencing methods:‌ long reads (ONT -‌​‌ PacBio) and linked reads​​ (Haplotagging) in two species,​​​‌ Coelopa frigida and Ciona‌ intestinalis. Analysing long-read‌​‌ high-coverage data was the​​ most exhaustive approach to​​​‌ detect and genotype inversions‌ of all lengths. It‌​‌ proved to be very​​ efficient to refine the​​​‌ breakpoints of adaptive inversions‌ previously detected by an‌​‌ indirect method. Yet, long-reads​​ cannot scale-up to population-wide​​​‌ studies in a cost-efficient‌ manner. Analysing linked-reads data‌​‌ thus provided a relevant​​ alternative to assess the​​​‌ position and frequency of‌ inversions across many more‌​‌ individuals. Overall, the combination​​ of both datasets provided​​​‌ insights into the distribution‌ of inversions across the‌​‌ genomes of both species​​ and across several populations​​​‌ 30, 29,‌ 53.

8.4.3 A‌​‌ review and roadmap for​​ the adoption of pangenomics​​​‌ in agronomy

Participants: Siegfried‌ Dubois, Fabrice Legeai‌​‌, Claire Lemaitre.​​

Pangenomics is transforming the​​​‌ way genomics is done.‌ By using assemblies of‌​‌ multiple complete genomes for​​ the same species, it​​​‌ is possible to depart‌ from a biased, reference-centric,‌​‌ point of view. Represented​​ as graphs, these pangenomes​​​‌ open new paths for‌ genome understanding and improvement‌​‌ of species of agricultural​​ interest. However, transitioning to​​​‌ a graph-based approach is‌ not straightforward, and many‌​‌ tools are still needed​​ for this shift. While​​​‌ numerous papers present both‌ methodological and applied approaches‌​‌ on pangenomics, and many​​ reviews present advantages of​​​‌ these methods, very few‌ resources outline what remains‌​‌ to be done. In​​ this opinion paper, we​​​‌ list the bottlenecks that‌ pangenome users experience once‌​‌ they have built their​​ graphs and the methods​​​‌ that are still required‌ to fully exploit pangenomes,‌​‌ and favor the widespread​​ adoption of this approach​​​‌43.

8.5.1‌​‌ Characterization of the newborn​​ microbiota and prediction of​​​‌ metabolite production

Participants: Lune‌ Angevin, Josipa Lipovac‌​‌, Pierre Peterlongo,​​ Emeline Roux, Riccardo​​​‌ Vicedomini.

Postnatal brain‌ development is influenced by‌​‌ the gut microbiota and​​ its metabolic activity, which​​​‌ are strongly shaped by‌ early-life diet. When breastfeeding‌​‌ is not possible, improving​​​‌ infant formula through targeted​ microbial supplementation requires precise​‌ identification of bacterial communities​​ and their functional potential,​​​‌ ideally at the strain​ level. This work focuses​‌ on the characterization of​​ the gut microbiota and​​​‌ the prediction of metabolite​ production in piglets fed​‌ with breast milk or​​ infant formula using Nanopore​​​‌ R10.4 sequencing of fecal​ samples 51, 28​‌. Long-read taxonomic assignment​​ tools were evaluated using​​​‌ publicly available real-data mock​ communities 27. Since​‌ functional diversity often occurs​​ at the strain level,​​​‌ the analysis concentrated on​ strain-level classifiers such as​‌ MADRe 49 and ORI​​ 98. MADRe is​​​‌ able to identify bacterial​ strains on very large​‌ reference databases but, at​​ the same time, it​​​‌ may produce a non-negligible​ number of false positives.​‌ In contrast, ORI yields​​ fewer false positives at​​​‌ the cost of significantly​ longer runtimes. To combine​‌ their strengths and cope​​ with large reference databases,​​​‌ we designed a pipeline​ in which MADRe provides​‌ a reduced set of​​ candidate genomes and ORI​​​‌ refines the selection. A​ final filtering step based​‌ on gene coverage is​​ then applied to remove​​​‌ remaining false positives. This​ approach makes it possible​‌ to reliably identify strains,​​ at the cost of​​​‌ losing some low-abundance strains.​ Future work will focus​‌ on reconstructing bacterial metabolic​​ networks and predicting the​​​‌ metabolites they may produce.​

8.5.2 Accurate MAG reconstruction​‌ from complex soil microbiome​​ through combined short- and​​​‌ HiFi long-reads metagenomics

Participants:​ Claire Lemaitre, Nicolas​‌ Maurice, Riccardo Vicedomini​​.

Advances in high-fidelity​​​‌ long-read (HiFi-LR) sequencing technologies​ offer unprecedented opportunities to​‌ uncover the microbial genomic​​ diversity of complex environments,​​​‌ such as soils. While​ short-read (SR) sequencing has​‌ enabled broad insights at​​ gene-level diversity, the inherently​​​‌ limited read length constrains​ the reconstruction of complete​‌ genomes. Conversely, HiFi-LR sequencing​​ enhances the quality and​​​‌ completeness of metagenome-assembled genomes​ (MAGs), enabling higher-resolution taxonomic​‌ and functional annotation. However,​​ the cost and relatively​​​‌ low throughput of HiFi-LR​ sequencing can limit genome​‌ recovery, particularly at the​​ binning stage, where coverage​​​‌ depth is critical. Here,​ we present a novel​‌ hybrid strategy that differs​​ from classical hybrid assemblies,​​​‌ where SR and LR​ reads are jointly used​‌ at the assembly step.​​ Instead, we use high-depth​​​‌ SR data to improve​ the binning of HiFi-LR​‌ contigs. Using both SR​​ and HiFi-LR metagenomic data​​​‌ generated from a tunnel-cultivated​ soil sample, we demonstrate​‌ that SR-derived coverage information​​ significantly improves the binning​​​‌ of HiFi-LR assemblies. This​ results in a substantial​‌ increase in the number​​ and quality of recovered​​​‌ MAGs compared to using​ HiFi-LR data alone and​‌ an incomparable improvement compared​​ to SR data alone.​​​‌ This approach underscores the​ potential of leveraging the​‌ vast amount of publicly​​ available Illumina metagenomic datasets.​​​‌ Completing existing SR resources​ with PacBio HiFi sequencing​‌ can maximise assembly contiguity​​ and binning accuracy using​​​‌ massive amounts of SR​ data already generated42​‌.

8.5.3 Novel genes​​ arising from genomic deletions​​​‌ across the bacterial tree​ of life

Participants: Karel​‌ Břinda.

Bacteria are​​ hosts to enormous genic​​ diversity. How new genes​​​‌ emerge, functionalize, and spread‌ remain longstanding questions. Here,‌​‌ we explore a mechanism​​ by which adaptive deletions​​​‌ fuse distant gene fragments.‌ Unlike other gene birth‌​‌ mechanisms that begin with​​ rare, neutral mutations, these​​​‌ “deletion-born fusions” reach high‌ frequency by hitch-hiking on‌​‌ the deletion. The deletion-driven​​ proliferation of the fusion​​​‌ prolongs the mutational supply‌ within these genes before‌​‌ loss, providing additional opportunities​​ for neofunctionalization. We document​​​‌ one such gene fixing‌ and expressing in a‌​‌ long-term E. coli evolution​​ experiment, and identify additional​​​‌ fusion events in the‌ Mycobacterium tuberculosis-bovis split. Finally,‌​‌ we develop a scalable​​ systematic screen to detect​​​‌ these genes in all‌ 2.4 million public single-isolate‌​‌ genomes and identify deletion-born​​ fusions across the bacterial​​​‌ tree of life. These‌ findings challenge the notion‌​‌ that deletions are solely​​ destructive and highlight their​​​‌ role as potential catalysts‌ for evolutionary innovation. 47‌​‌.

8.5.4 Diversity of​​ genomic structural variation across​​​‌ the Tree of Life‌

Participants: Claire Lemaitre.‌​‌

Understanding why some species​​ have higher levels of​​​‌ genetic diversity than others‌ is central in ecology,‌​‌ and has important implications​​ for the evolution and​​​‌ conservation of species. Current‌ research in evolutionary genomics‌​‌ is largely based on​​ Single Nucleotide Polymorphism (SNP),​​​‌ therefore neglecting the genetic‌ diversity and complexity introduced‌​‌ by Structural Variants (SVs)​​ within populations and across​​​‌ species. Leveraging the Darwin‌ Tree of Life project,‌​‌ which routinely generates PacBio​​ HiFi long reads and​​​‌ high-quality genome assemblies for‌ a broad range of‌​‌ taxa, we aim to​​ thoroughly characterize SV diversity​​​‌ across the tree of‌ life. To achieve this,‌​‌ we developed a Snakemake​​ pipeline for detecting heterozygote​​​‌ SVs, integrating multiple aligners‌ and SV callers to‌​‌ retain only the highest-quality​​ SVs for downstream comparative​​​‌ analysis55, 56‌.

Besides the necessary‌​‌ and valuable description of​​ SV distribution and determinants​​​‌ on a large phylogenetic‌ scale, this comprehensive dataset‌​‌ allows us to address​​ key evolutionary questions through​​​‌ the lens of structural‌ variation, such as understanding‌​‌ the variability of (structural)​​ genetic diversity across species​​​‌ and examining the overall‌ role of SVs in‌​‌ evolutionary processes. For example,​​ we investigate how variation​​​‌ in effective population size‌ (Ne) between species drives‌​‌ the distribution of structural​​ genetic diversity at a​​​‌ macro-evolutionary scale. We also‌ test whether and how‌​‌ genetic diversity scales with​​ the distribution of different​​​‌ types of genetic variants‌ including different types of‌​‌ SVs, indels, and SNPs.​​ Genetic diversity also appears​​​‌ to be predicted by‌ the species’ biology and‌​‌ ecology. We thus compare​​ species with different life​​​‌ histories (such as longevity‌ and body size), as‌​‌ well as variations in​​ geographic range and population​​​‌ census size, to test‌ whether these factors also‌​‌ apply to SVs. Overall,​​ our integrative analysis, combining​​​‌ genomic and ecological data,‌ should be key to‌​‌ understanding the variability of​​ structural genetic diversity across​​​‌ species and examining the‌ role of SVs in‌​‌ evolutionary processes54.​​

8.5.5 The Silene latifolia​​​‌ genome and its giant‌ Y chromosome

Participants: Claire‌​‌ Lemaitre.

In some​​​‌ species, the Y is​ a tiny chromosome but​‌ the dioecious plant Silene​​ latifolia has a giant​​​‌ Y chromosome of 550​ Mb, which has remained​‌ unsequenced so far. We​​ participated in the assembly​​​‌ and comparative analysis of​ a high-quality male S.​‌ latifolia genome. Comparative analysis​​ of the sex chromosomes​​​‌ with outgroups showed the​ Y is surprisingly rearranged​‌ and degenerated for a​​ 11 MY-old system. Recombination​​​‌ suppression between X and​ Y extended in a​‌ stepwise process, and triggered​​ a massive accumulation of​​​‌ repeats on the Y,​ as well as in​‌ the non-recombining pericentromeric region​​ of the X, leading​​​‌ to giant sex chromosomes​ 16.

8.5.6 Genomics​‌ and transcriptomics of Brassicaceae​​ plants and agro-ecosystem insects​​​‌

Participants: Fabrice Legeai.​

Through its long-term collaboration​‌ with INRAE IGEPP, and​​ its support to the​​​‌ BioInformatics of Agroecosystems Arthropods​ platform, GenScale is​‌ involved in various genomic​​ and transcriptomics projects in​​​‌ the field of agricultural​ research, and participated in​‌ the genome assembly and​​ analyses of some major​​​‌ agricultural pests or their​ natural enemies. In particular,​‌ we perfomed a comparative​​ analysis of olfactory (OR)​​​‌ and gustatory receptor (GR)​ genes and transposable elements​‌ (TE) across 12 aphid​​ genomes with varying host​​​‌ ranges17. Our​ results suggest that TE​‌ activity may facilitate functional​​ innovation in GRs while​​​‌ alleviating constraints or pseudogenization​ in ORs. This study​‌ reveals how duplication, selection,​​ and TE dynamics shape​​​‌ gene evolution in insect​ pests. It also provides​‌ the first chromosome-scale genome​​ assembly of Dysaphis plantaginea​​​‌, with comprehensive annotations​ and functional predictions of​‌ OR/GR genes, bridging adaptive​​ evolution with mechanistic insights.​​​‌ We also participated to​ the annotation of the​‌ large lepidoptera genome of​​ Hylesia metabus, and​​​‌ the comparison of its​ content with 17 additional​‌ Saturniidae and Sphingidae genomes​​ suggesting that an accumulation​​​‌ of repetitive elements likely​ led to the increased​‌ size of its genome​​ 18. Using a​​​‌ transcriptomics approach, we identified​ genes differentially expressed between​‌ gonadal and non‑gonadal tissues​​ of Spodoptera frugiperda,​​​‌ this study targetted kinetochore​ genes significantly overexpressed in​‌ the gonads than in​​ somatic tissues 13.​​​‌ Furthermore, we compared the​ genomes of 46 aphid​‌ species, and identified genes​​ evolving faster in species​​​‌ that do not alternate​ their life cycle (i.e.​‌ migrating between highly distinct​​ plant hosts), suggesting that​​​‌ the loss of a​ complex life cycle leads​‌ to reduced selective constraints​​ as a consequence of​​​‌ ecological simplification 20.​ We also helped to​‌ characterize the presence of​​ Pseudomonas brassicacearum harboring a​​​‌ gene able to detoxifying​ harmful isothiocyanates in various​‌ Brassica napus genotypes, as​​ well as in the​​​‌ guts of the herbivore​ Delia radicum. As​‌ we observed that this​​ bacteria was shared consistently​​​‌ in plant and insect,​ we demonstrated that plant​‌ genotypes can shape the​​ gut microbial communities of​​​‌ its pests by promoting​ the acquisition of symbiotic​‌ bacteria 9. Finally,​​ we also participated to​​​‌ the production of important​ crop genomics resources such​‌ as three Brassica oleracea​​ and two Brassica rapa​​12, or the​​​‌ assembly of Dactylis glomerata‌ and Medicago sativa.‌​‌ These two last genomes​​ are particularly difficult to​​​‌ assemble due to their‌ polyploidy (auto tetraploid) and‌​‌ their large sizes (3.2​​ and 7.5 Gbp respectively)​​​‌ 23.

8.5.7 Prediction‌ of genetic relatedness of‌​‌ Escherichia coli using neighbour​​ typing: A tool for​​​‌ rapid outbreak detection

Participants:‌ Karel Břinda.

Identifying‌​‌ the genetic relatedness of​​ resistant bacterial pathogens in​​​‌ healthcare settings can help‌ identify undetected transmission events‌​‌ and outbreaks. However, current​​ methods are time- and​​​‌ resource-intensive. We evaluated a‌ rapid neighbour typing method‌​‌ paired with long-read sequencing​​ for assessment of genetic​​​‌ relatedness.Utilizing a dataset of‌ primary clinical samples and‌​‌ published isolate data from​​ two outbreaks of Escherichia​​​‌ coli, we applied genomic‌ neighbour typing of long-‌​‌ read sequence data to​​ rapidly estimate genetic relatedness.​​​‌ We assessed the correlation‌ between neighbour typing predicted‌​‌ genetic distance and pairwise​​ genetic distance from short-read​​​‌ draft whole genomes for‌ all sample pairs. Predicted‌​‌ genetic trees using neighbour​​ typing were compared to​​​‌ reference genetic trees generated‌ using mash distances and‌​‌ maximum-likelihood (ML) methods to​​ assess the extent of​​​‌ agreement, along with metrics‌ of cluster similarity (cluster‌​‌ comparability and Baker’s gamma​​ index) and tree topology​​​‌ similarity (generalized Robinson-Foulds metric).For‌ all three datasets, we‌​‌ found strong correlations between​​ the reference methods and​​​‌ predicted genetic distances (Spearman’s‌ rho=0.75-0.95, p<0.001), which improved‌​‌ when using a lineage​​ score informed approach (Spearman’s​​​‌ rho=0.93-0.94, p<0.001). Predicted genetic‌ trees and clusters from‌​‌ neighbour typing were comparable​​ to those generated using​​​‌ either mashtree or a‌ ML method, with a‌​‌ range of cluster comparability​​ of 85.8%-99.5%, Baker’s gamma​​​‌ indices of 0.8-0.95, and‌ generalized Robinson-Foulds values of‌​‌ 0.34-0.8.Pairing the neighbour typing​​ method with long-read sequencing​​​‌ can enable accurate predictions‌ of the relatedness of‌​‌ E. coli samples and​​ isolates, and could potentially​​​‌ be used as a‌ rapid outbreak surveillance tool.‌​‌ 10, 22

8.5.8​​ Neighbour Typing Using Long​​​‌ Read Sequencing Provides Rapid‌ Prediction of Sequence Type‌​‌ and Antimicrobial Susceptibility of​​ Klebsiella pneumoniae

Participants: Karel​​​‌ Břinda.

The rapid‌ genome-based diagnostic approach of‌​‌ long read sequencing coupled​​ with neighbor typing offers​​​‌ the potential to improve‌ empiric treatment of infection.‌​‌ However, this approach is​​ still in development, and​​​‌ clinical validation is needed‌ to support its use.‌​‌ In this study, we​​ present an assessment of​​​‌ a neighbour typing method‌ (RASE - resistance associated‌​‌ sequence elements) to predict​​ lineage or sequence type​​​‌ (ST) and antimicrobial susceptibility‌ in real time for‌​‌ Klebsiella pneumoniae sensu lato.​​ We analysed the initial​​​‌ reads generated during the‌ early phase of long‌​‌ read sequencing from pure​​ culture (n=99), mock communities​​​‌ (n=20) and metagenomic samples‌ (n=20). RASE accurately identified‌​‌ 69.7% and 70% of​​ STs in pure culture​​​‌ and metagenomes, respectively, and‌ identified the STs of‌​‌ the isolates representing the​​ highest proportion in mock​​​‌ communities. Regarding antimicrobial susceptibility‌ prediction, the probability of‌​‌ susceptibility increased to 72%​​ (95% CI 63%-80%) across​​​‌ all tested antibiotics, when‌ RASE predicted susceptibility, and‌​‌ decreased probability of susceptibility​​​‌ to 8.9% (95% CI​ 6.4%-9.6%) when was indicative​‌ of a resistant phenotype.​​ Our study confirmed that​​​‌ genomic neighbor typing in​ K. pneumoniae sensu lato​‌ is capable of providing​​ informative predictions of ST​​​‌ and antibiotic susceptibility in​ less than ten minutes​‌ (after the start of​​ sequencing) with 200-500 reads.​​​‌ IMPORTANCE The growing burden​ of antimicrobial resistance is​‌ leading to high rates​​ of mortality and morbidity​​​‌ worldwide. This situation has​ made the selection of​‌ empirical antibiotic therapy challenging,​​ due to the risk​​​‌ of treatment failure and​ the overuse of last-resort​‌ antibiotics. The development of​​ new sequencing technologies is​​​‌ helping to reduce the​ waiting time for a​‌ microbiological diagnosis, providing information​​ in the early phase​​​‌ of bacterial infections, which​ could help improve clinical​‌ outcomes in a time​​ of rising antimicrobial resistance.​​​‌ In this context, we​ assessed the performance of​‌ RASE (resistance associated sequence​​ elements) in Klebsiella pneumoniae​​​‌ , an opportunistic pathogen​ frequently associated with nosocomial​‌ infections, which can rapidly​​ acquire antibiotic resistance genes.​​​‌ Thus, in our study​ we provide insights that​‌ may aid in the​​ validation of RASE for​​​‌ clinical use 45.​ The work was also​‌ presented as a poster​​ at Applied Bioinformatics and​​​‌ Public Health Microbiology 2025​ 57.

10.1.1 Participation in​​ other International Programs

10.2.1​ Visits of international scientists​‌

10.2.2 Visits to​‌ international teams

10.3.1 Other​​ european programs/initiatives

10.4.1 PEPR

10.4.2​​ ANR

10.4.3​ Inria Exploratory Action

Rennes Metropole: Allocation d'installation​‌ scientifique

Participants: Karel Břinda​​.

• Coordinator: Karel Brinda​​
• Duration: 20 months (2024-2026)​​​‌
• Description: Fast DNA sequence‌ data search is crucial‌​‌ for our ability to​​ control the spread of​​​‌ infectious diseases. However, it‌ represents a significant challenge‌​‌ in data science: exponentially​​ growing sequencing data exceeds​​​‌ the development of computing‌ capabilities. The central goal‌​‌ of this project is​​ to develop sublinear algorithms​​​‌ for searching within collections‌ of bacterial genomes, with‌​‌ the ultimate aim of​​ enabling searches on all​​​‌ sequenced bacteria on standard‌ desktop and laptop computers.‌​‌ This will be achieved​​ through advancements in phylogenetic​​​‌ compression and entropy scale‌ algorithms.

11.1.1‌​‌ Scientific events: organisation

11.1.2​​​‌ Scientific events: selection

11.1.3​​​‌ Journal

11.1.4 Invited talks

• "From‌​‌ bacteria to bits and​​ back again: faster, more​​​‌ accurate, and smarter diagnostics‌ of antibiotic resistance", Czech-French‌​‌ Science Meetup 2025 at​​ the Czech Embassy in​​​‌ Paris, Oct. 2025 [‌Karel Břinda ]
• "Moving‌​‌ from reference genomes to​​ pangenome graphs: benefits and​​​‌ challenges for the analysis‌ of Structural Variation", BiGre‌​‌ Days: Computational biology seminars​​ in Grenoble, Grenoble,​​​‌ Feb. 2025 [Claire‌ Lemaitre ]
• "Archivage de‌​‌ données numériques sur ADN",​​ Séminaire des Archives Départementales​​​‌ d'Ille et Vilaine, Rennes,‌ Dec. 2025 [Dominique‌​‌ Lavenier ]
• "Open challenges​​​‌ in pangenomics", at "Biology​ In Silico" Meetings of​‌ IGDR Rennes, Apr. 2025.​​ [Siegfried Dubois ]​​​‌
• "Moving from reference genomes​ to pangenome graphs: benefits​‌ and challenges for the​​ analysis of Structural Variation",​​​‌ at "Biology In Silico"​ Meetings of IGDR Rennes,​‌ Apr. 2025. [Claire​​ Lemaitre ]
• "Generative AI:​​​‌ worth the cost?", AI4Dev,​ Grenoble, Feb. 2025. [​‌Siegfried Dubois ]
• "Toward​​ sublimera algorithms for searching​​​‌ large genome databases", at​ Pavel Veselý's group at​‌ Charles University, Czech Republic,​​ Dec. 2025 [Leo​​​‌ Ackermann ]

11.1.5 Leadership​ within the scientific community​‌

• Member of the Scientific​​ Advisory Board of the​​​‌ GDR BIMMM (National Research​ Group in Molecular Bioinformatics)​‌ [Claire Lemaitre ]​​
• Animator of the Sequence​​​‌ Algorithms axis (seqBIM​ GT) of the​‌ GDRs BIMMMM and IFM​​ (National Research Groups in​​​‌ Molecular Bioinformatics and Informatics​ and Mathematics respectively) (350​‌ French participants) [Claire​​ Lemaitre ]
• Member of​​​‌ the PEPR MoleculArxiv Executive​ Committee [Dominique Lavenier​‌ ]

11.1.6 Scientific expertise​​

• Expert at DAEI (Pole​​​‌ expertise international de la​ Délégation au Affaire Européennes​‌ et Internationales), MESR [​​Dominique Lavenier ]

11.1.7​​​‌ Research administration

• Corresponding member​ of COERLE (Inria Operational​‌ Committee for the assessment​​ of Legal and Ethical​​​‌ risks) [Jacques Nicolas​ ]
• Recruitment committees: member​‌ of a Univ Rennes​​ CPJ selection committee [​​​‌Claire Lemaitre ]
• Scientific​ committees: member of the​‌ scientific committee of the​​ INRAE's Plant Health and​​​‌ Environment department [Fabrice​ Legeai ]
• Strategic data​‌ referent for the INRAE's​​ Plant Health and Environment​​​‌ department [Fabrice Legeai​ ]

11.2.1 Teaching administration

• Head​​ of the master's degree​​​‌ "Nutrition Sciences des Aliments"​ (NSA) of University of​‌ Rennes (75 students) [​​Emeline Roux ]

11.2.2​​​‌ Teaching

• Licence: Linear Programming,​ 30h, L3 Miage, Univ.​‌ Rennes, France [Roumen​​ Andonov ]
• Licence :​​​‌ Biochemistry, 90h, L1 and​ L3, Univ. Rennes, France​‌ [Emeline Roux ]​​
• Licence: Object Oriented Programming,​​​‌ 38h, L2, INSA Rennes,​ France [Nicolas Maurice​‌ ]
• Licence: Introduction to​​ Data Analysis, 28h, L2​​​‌ Informatics, Univ. Rennes, France​ [Siegfried Dubois ]​‌
• Licence: Object-Oriented Programming, 28h,​​ L2, Univ. Rennes, France​​​‌ [Alix Regnier ]​
• Master: Short Introduction to​‌ R,22h30, M1 Bioinformatics, Univ.​​ Rennes, France [Mélody​​​‌ Temperville ]
• Master: Data​ Visualisation,9h, M1 Bioinformatics, Univ.​‌ Rennes, France [Melody​​ Temperville ]
• Master: General​​​‌ Statistics, 33h, M1 Nutrition​ and Food Science, Univ.​‌ Rennes, France [Lune​​ Angevin ]
• Master: Experimental​​​‌ Bioinformactics, 24h, M1, ENS​ Rennes, France [Léo​‌ Ackermann , Karel Břinda​​ ]
• Master: Sequence Bioinformatics,​​​‌ 42h, M1 Informatics, Univ.​ Rennes, France [Claire​‌ Lemaitre , Pierre Peterlongo​​ , Alix Regnier ]​​​‌
• Master: Algorithms on Sequences,​ 52h, M2 Bioinformatics, Univ.​‌ Rennes, France [Léo​​ Ackermann , Claire Lemaitre​​ , Pierre Peterlongo ]​​​‌
• Master : Biochemistry and‌ microbiology, 130h, M1 and‌​‌ M2, Univ. Rennes, France​​ [Emeline Roux ]​​​‌
• Master : Bioanalysis, 6h,‌ M2 Institut Agro Rennes‌​‌ Angers, France [Fabrice​​ Legeai ]

11.2.3 Supervision​​​‌

• PhD: Meven Mognol ,‌ Processing-in-Memory 41, CIFRE,‌​‌ defense: July 2025, [​​Dominique Lavenier ]
• PhD​​​‌ in progress: Siegfried Dubois‌ , Characterizing structural variation‌​‌ in pangenome graphs, Inria​​ (PEPR-ANR Agrodiv), since 15/09/2023,​​​‌ [Claire Lemaitre ,‌ T. Faraut, M. Zynticki.]‌​‌
• PhD in progress: Nicolas​​ Maurice , Sequence algorithmics​​​‌ for de novo genome‌ assembly from complex metagenomic‌​‌ data, Inria (PEPR-ANR Mistic),​​ since 01/10/2023, [Claire​​​‌ Lemaitre , C. Frioux,‌ Riccardo Vicedomini ].
• PhD‌​‌ in progress: V. Levallois,​​ Indexing genomic data, Défi​​​‌ Inria OmicFinder, since 01/10/2023,‌ [Pierre Peterlongo ].‌​‌
• PhD in progress: Y.​​ Tirlet (IRISA Dyliss team),​​​‌ Univ Rennes (ANR), Integrative‌ method for multi-omics data‌​‌ analysis with application to​​ the activation and regulation​​​‌ of an endogeneized viral‌ genome in a parasitoid‌​‌ wasp, since 01/05/2023, [E.​​ Becker, O. Dameron, Fabrice​​​‌ Legeai .]
• PhD in‌ progress: Francesca Brunetti (Sapienza‌​‌ University of Rome), Assessment​​ of genetic determinants in​​​‌ Escherichia coli uropathogenic lifestyle‌ and intracellular persistence via‌​‌ optimized k-mer matching of​​ million-genome collections on laptops,​​​‌ since 01/11/2022, [Karel‌ Břinda , Maria Pia‌​‌ Conte]
• PhD in progress:​​ Alix Regnier , Limiting​​​‌ the size of data‌ structures indexing genomic sequences,‌​‌ Inria (PEPR-ANR Agrodiv), since​​ 01/10/2024, [Pierre Peterlongo​​​‌ ].
• PhD in progress:‌ Léo Ackermann , Developing‌​‌ efficient algorithms for sublinear​​ search in large genome​​​‌ databases, Inria, since 01/10/2024,‌ [Karel Břinda ,‌​‌ Pierre Peterlongo ].
• PhD​​ in progress: T. Truong,​​​‌ Computational methods for phylogenetic‌ compression, Univ Rennes, since‌​‌ 01/11/2024, [Karel Břinda​​ , Pierre Peterlongo ,​​​‌ Dominique Lavenier ].
• PhD‌ in progress: Lune Angevin‌​‌ , Fine characterization of​​ the intestinal microbiota and​​​‌ prediction of metabolite production,‌ Univ Rennes, since 01/10/2024,‌​‌ [Emeline Roux ,​​ Riccardo Vicedomini , Pierre​​​‌ Peterlongo ].
• PhD in‌ progress: M. Temperville, Methods‌​‌ for characterizing structural variations​​ in genomes with linked-read​​​‌ data, Univ Rennes, since‌ since 01/10/2024, [Fabrice‌​‌ Legeai , Claire Lemaitre​​ , Claire Mérot].

11.2.4​​​‌ Juries

• Reviewer of HDR‌ thesis
  • Antoine Limasset, Univ.‌​‌ Lille, Sept. 2025 [​​Claire Lemaitre ]
  • Raja​​​‌ Appuswamy, Sorbone University, Oct.‌ 2025 [Dominique Lavenier‌​‌ ]
• Member of HDR​​ thesis jury
  • Raluca Uricaru,​​​‌ Bordeaux University, Fev. 2025‌ [Pierre Peterlongo ]‌​‌
• President of PhD thesis​​ jury
  • Francesco Andreace, Pasteur​​​‌ institute Paris, Jan. 2025,‌ [Pierre Peterlongo ]‌​‌
• Reviewer of PhD thesis​​
  • Rick Wertenbroek, Univ. Lausanne,​​​‌ May 2025 [Dominique‌ Lavenier ]
  • Thomas Baudeau,‌​‌ Univ. Lille, July 2025​​ [Dominique Lavenier ]​​​‌
  • Lea Vandamme, Univ. Lille,‌ Dec. 2025 [Dominique‌​‌ Lavenier ]
  • Lucas Parmigiani,​​ Bielefeld University, Germany, Nov​​​‌ 2025 [Pierre Peterlongo‌ ]
• Member of PhD‌​‌ thesis jury
  • Mael Lefeuvre,​​ Natural history museum Paris,​​​‌ Dec 2025 [Pierre‌ Peterlongo ]
  • Nastasija Mijovic,‌​‌ Univ. Montpellier, Dec 2025​​ [Pierre Peterlongo ]​​​‌
• Member of PhD thesis‌ committee
  • Léa Nicolas, Univ‌​‌ Rennes [Claire Lemaitre​​​‌ ]
  • Soraya Belbati, Univ​ Rennes [Claire Lemaitre​‌ ]
  • Dimple Adiwal, Univ​​ Rennes [Claire Lemaitre​​​‌ ]
  • Jules Burgat, Univ​ Rennes [Claire Lemaitre​‌ ]
  • Florent Couturier, Univ​​ Bordeaux [Claire Lemaitre​​​‌ ]
  • O. Weppe, Univ.​ Rennes [P. Peterlongo]
  • Silpadas​‌ Nedoolil, Univ. Rennes [P.​​ Peterlongo]
  • Mael Coatanhay, Univ.​​​‌ Rennes [P. Peterlongo]
  • H.​ Gasnier, IMT-A, Univ. Brest​‌ [Dominique Lavenier ]​​
  • F. Santoro, UBS, Lorient​​​‌ [Dominique Lavenier ]​
  • M. Caneve, IMT-A, Univ.​‌ Brest [Dominique Lavenier​​ ]
  • Raneem Jaafar, Univ.​​​‌ Rennes [Riccardo Vicedomini​ ]
  • Fantine Benoit, Univ.​‌ Rennes [Riccardo Vicedomini​​ ]
  • Emile Breton, Univ.​​​‌ Rennes [Karel Břinda​ ]
  • Alexandra Jalaber Dupont​‌ de Dinechin, Univ. Paris​​ Saclay [Fabrice Legeai​​​‌ ]

11.3.1 Specific official​​ responsibilities in science outreach​​​‌ structures

• Member of the​ Interstice editorial board [​‌Pierre Peterlongo ]

11.3.2​​ Productions (articles, videos, podcasts,​​​‌ serious games, ...)

• Media​ coverage: Mieux caractériser l’antibiorésistance.​‌ Sciences Ouest (2025/9, N◦431),​​ a popularization article about​​​‌ our research on antibiotic​ resistance [Karel Břinda​‌ ]
• Article: Diagnostiquer plus​​ vite la résistance aux​​​‌ antibiotiques, Inria Emergences, a​ popularization article about our​‌ research on antibiotic resistance​​ [Karel Břinda ]​​​‌ [web link]​
• Video: DNAmaker: Automation of​‌ the design of long​​ molecules [Julien Leblanc​​​‌ ]. [youtube link​]
• Polularization book: Stocker​‌ nos données sur ADN​​ [Dominique Lavenier ]​​​‌ 6160
• Thès'en Images​ Presentation of PhD and​‌ a thesis project to​​ a group of young​​​‌ people who had dropped​ out of school at​‌ AFPA, which enabled them​​ to make a video​​​‌ with the association Thès'en​ Images. One intervention in​‌ 2025. [Lune Angevin​​ ]
• Article: The conversation​​​‌ Biodiversité alimentaire, microbiote et​ bien-être : la recherche​‌ explore les liens potentiels.​​ [Emeline Roux ]​​​‌ 62
• Game: Unplugged activity:​ Form'IAdable (projet TIARe), a​‌ game about supervised learning​​ and its biases [​​​‌Leo Ackermann ] [​web link]

11.3.3​‌ Participation in Live events​​

• Scientific animator at la​​​‌ fête de la Science​ aux Champs Libres,​‌ Oct. 2025, Rennes [​​Siegfried Dubois ]
• Sciences​​​‌ à la Une Creation​ of a research project​‌ for a sixth grade​​ classroom leading to the​​​‌ production of a scientific​ podcast. Several interventions in​‌ 2025. [Lune Angevin​​ ]
• Rendez-vous des Jeunes​​​‌ Mathématiciennes et Informaticiennes Discussion​ with female high school​‌ students about university studies,​​ computer science and mathematics.​​​‌ One intervention in 2025.​ [Lune Angevin ]​‌

International journals‌

• 7 articleC.Chloé‌​‌ Berton, G.Gouenou​​ Coatrieux, H.Héléne​​​‌ Gasnier, S.Sahar‌ Haddad, R.Reda‌​‌ Bellafqira and D.Dominique​​ Lavenier. A Dynamic​​​‌ Sliding Window Encoding for‌ Secured DNA Data Storage‌​‌ Compliant With Biological and​​ Indexing Constraints.IEEE​​​‌ Access132025,‌ 66009-66030HALDOIback‌​‌ to text
• 8 article​​K.Karel Břinda,​​​‌ L.Leandro Lima,‌ S.Simone Pignotti,‌​‌ N.Natalia Quinones-Olvera,​​ K.Kamil Salikhov,​​​‌ R.Rayan Chikhi,‌ G.Gregory Kucherov,‌​‌ Z.Zamin Iqbal and​​ M.Michael Baym.​​​‌ Efficient and robust search‌ of microbial genomes via‌​‌ phylogenetic compression.Nature​​ Methods22April 2025​​​‌, 692–697HALDOI‌back to textback‌​‌ to text
• 9 article​​​‌J. M.Juliette M.​ Carpentier, S.Stéphane​‌ Derocles, S.Sylvain​​ Chéreau, B.Bruno​​​‌ Marquer, J.Juliette​ Linglin, L.Lionel​‌ Lebreton, F.Fabrice​​ Legeai, N.Nathan​​​‌ Vannier, A. M.​Anne Marie Cortesero and​‌ C.Christophe Mougel.​​ Contrasting glucosinolate profiles in​​​‌ rapeseed genotypes shape the​ rhizosphere-insect continuum and microbial​‌ detoxification potential in a​​ root herbivore.Msystems​​​‌10122025,​ e01269-25HALDOIback​‌ to text
• 10 article​​A. C.Amanda C.​​​‌ Carroll, L.Leanne​ Mortimer, H.Hiren​‌ Ghosh, S.Sandra​​ Reuter, H.Hajo​​​‌ Grundmannd, K.Karel​ Břinda, W. P.​‌William P. Hanage,​​ A.Angel Li,​​​‌ A.Aimee Paterson,​ A.Andrew Purssell,​‌ A. M.Ashley M.​​ Rooney, N. R.​​​‌Noelle R. Yee,​ B.Bryan Coburn,​‌ S.Shola Able-Thomas,​​ M.Martin Antonio,​​​‌ A.Allison Mcgeerg and​ D. R.Derek R.​‌ Macfadden. Prediction of​​ genetic relatedness of Escherichia​​​‌ coli using neighbour typing:​ A tool for rapid​‌ outbreak detection.Antimicrobial​​ Agents and Chemotherapy2025​​​‌. In press. HAL​back to text
• 11​‌ articleS.Siegfried Dubois​​, M.Matthias Zytnicki​​​‌, C.Claire Lemaitre​ and T.Thomas Faraut​‌. Pairwise graph edit​​ distance characterizes the impact​​​‌ of the construction method​ on pangenome graphs.​‌Bioinformatics416June​​ 2025, btaf291HAL​​​‌DOIback to text​
• 12 articleC.Cyril​‌ Falentin, W.William​​ J.W. Thomas, M.​​​‌Matéo Boudet, P.​Pauline Le Boulch,​‌ A.Alexis Bourdais,​​ L.Loeiz Maillet,​​​‌ F.Fabrice Legeai,​ K.Karine Labadie,​‌ C.Corinne Cruaud,​​ G.Gwenaëlle Deniot,​​​‌ J.Jacqueline Batley,​ A.Antoine Gravot,​‌ J.-M.Jean-Marc Aury and​​ M.Mathieu Rousseau-Gueutin.​​​‌ Chromosome level assembly of​ five Brassica rapa and​‌ oleracea accessions expand the​​ resistance genes reservoir.​​​‌Scientific Data 121​2025, 2016HAL​‌DOIback to text​​
• 13 articleS.Sylvie​​​‌ Gimenez, M.Magali​ Eychenne, F.Fabrice​‌ Legeai, S.Sally​​ Gamble and E.Emmanuelle​​​‌ d'Alençon. Towards identification​ of a holocentromere marker​‌ in the lepidopteran model​​ <i>Spodoptera frugiperda</i>.Chromosoma​​​‌ Biology of the Nucleus​1341March 2025​‌, 2HALDOI​​back to text
• 14​​​‌ articleN.Nicolas Maurice​, C.Claire Lemaitre​‌, R.Riccardo Vicedomini​​ and C.Clémence Frioux​​​‌. Mapler: a pipeline​ for assessing assembly quality​‌ in taxonomically rich metagenomes​​ sequenced with HiFi reads​​​‌.Bioinformatics416​June 2025, btaf334​‌HALDOIback to​​ text
• 15 articleF.​​​‌Florestan de Moor,​ O.Olivier Boullé and​‌ D.Dominique Lavenier.​​ De-Bruijn graph partitioning for​​​‌ scalable and accurate DNA​ storage processing.Bioinformatics​‌4111November 2025​​, 1-9HALDOI​​​‌back to text
• 16​ articleC.Carol Moraga​‌, C.Catarina Branco​​, Q.Quentin Rougemont​​​‌, P.Paris Veltsos​, P.Pavel Jedlička​‌, A.Aline Muyle​​, M.Mélissa Hanique​​, E.Eric Tannier​​​‌, X.Xiaodong Liu‌, E.Eddy Mendoza-Galindo‌​‌, C.Claire Lemaitre​​, P. D.Peter​​​‌ D. Fields, C.‌Corinne Cruaud, K.‌​‌Karine Labadie, C.​​Caroline Belser, J.​​​‌Jérôme Briolay, S.‌Sylvain Santoni, R.‌​‌Radim Cegan, R.​​Raquel Linheiro, R.​​​‌ C.Ricardo C. Rodríguez‌ de la Vega,‌​‌ G.Gabriele Adam,​​ A. E.Adil El​​​‌ Filali, V.Vinciane‌ Mossion, A.Adnane‌​‌ Boualem, R.Raquel​​ Tavares, A.Amine​​​‌ Chebbi, R.Richard‌ Cordaux, C.Cécile‌​‌ Fruchard, D.Djivan​​ Prentout, A.Amandine​​​‌ Velt, B.Bruno‌ Spataro, S.Stephane‌​‌ Delmotte, L.Laura​​ Weingartner, H.Helena​​​‌ Toegelová, Z.Zuzana‌ Tulpová, P.Petr‌​‌ Cápal, H.Hana​​ Šimková, H.Helena​​​‌ Štorchová, M.Manuela‌ Krüger, O. A.‌​‌Oushadee A. J. Abeyawardana​​, D. R.Douglas​​​‌ R. Taylor, M.‌ S.Matthew S. Olson‌​‌, D. B.Daniel​​ B. Sloan, S.​​​‌Sophie Karrenberg, L.‌ F.Lynda F. Delph‌​‌, D.Deborah Charlesworth​​, T.Tatiana Giraud​​​‌, A.Abdelhafid Bendahmane‌, A. D.Alex‌​‌ Di Genova, M.​​Mohammed‐Amin Madoui, R.​​​‌Roman Hobza and G.‌Gabriel Marais. The‌​‌ Silene latifolia genome and​​ its giant Y chromosome​​​‌.Science3876734‌February 2025, 630-636‌​‌HALDOIback to​​ text
• 17 articleS.​​​‌ G.Sergio Gabriel Olvera-Vazquez‌, X.Xilong Chen‌​‌, A.Aurélie Mesnil​​, C.Camille Meslin​​​‌, F.Fabricio Almeida-Silva‌, J.Johann Confais‌​‌, Y.Yann Bourgeois​​, G.Gianluca Lombardi​​​‌, C.Célia Lougmani‌, K.Karine Alix‌​‌, N.Nicolas Francillonne​​, N.Nathalie Choisne​​​‌, S.Stephane Cauet‌, J.-C.Jean-Christophe Simon‌​‌, C.Christelle Buchard​​, N.Nathalie Rodde​​​‌, D.David Ogereau‌, C.Claire Mottet‌​‌, A.Alexandre Degrave​​, E.Elorri Segura​​​‌, A.Alessandra Carbone‌, B.Barrès Benoît‌​‌, E.Emmanuelle Jacquin-Joly​​, W.William Marande​​​‌, D.Dominique Lavenier‌, F.Fabrice Legeai‌​‌ and A.Amandine Cornille​​. Comprehensive annotation of​​​‌ olfactory and gustatory receptor‌ genes and transposable elements‌​‌ revealed their evolutionary dynamics​​ in aphids.Molecular​​​‌ Biology and Evolution42‌12September 2025,‌​‌ msaf238HALDOIback​​ to text
• 18 article​​​‌C.Charles Perrier,‌ R.Rémi Allio,‌​‌ F.Fabrice Legeai,​​ M.Mathieu Gautier,​​​‌ F.Frédéric Bénéluz,‌ W.William Marande,‌​‌ A.Anthony Théron,​​ N.Nathalie Rodde,​​​‌ M.Melfran Herrera,‌ L.Laure Sauné,‌​‌ H.Hugues Parrinello,​​ M.Mélanie Mcclure and​​​‌ M.Mónica Arias.‌ Transposable element accumulation drives‌​‌ genome size increase in​​ Hylesia metabus (Lepidoptera: Saturniidae),​​​‌ an urticating moth species‌ from South America.‌​‌Journal of Heredity116​​3May 2025,​​​‌ 344-353HALDOIback‌ to text
• 19 article‌​‌O.Ondřej Sladký,​​ P.Pavel Veselý and​​​‌ K.Karel Břinda.‌ FroM Superstring to Indexing:‌​‌ a space-efficient index for​​​‌ unconstrained k-mer sets using​ the Masked Burrows-Wheeler Transform​‌ (MBWT).Bioinformatics Advances​​November 2025, 1-10​​​‌HALDOIback to​ text
• 20 articleT.​‌Théo Vericel, G.​​Gaorui Gong, F.​​​‌Fabrice Legeai, A.​Aurelie Etier, J.​‌Julie Jaquiéry and J.-C.​​Jean-Christophe Simon. Life​​​‌ Cycle Simplifications in Aphids​ Drive Changes in Evolutionary​‌ Rates and Selection Regimes​​.Molecular Biology and​​​‌ Evolution4212December​ 2025, msaf307HAL​‌DOIback to text​​
• 21 articleR.Riccardo​​​‌ Vicedomini, F.Francesco​ Andreace, Y.Yoann​‌ Dufresne, R.Rayan​​ Chikhi and C.Camila​​​‌ Duitama González. MUSET:​ set of utilities for​‌ constructing abundance unitig matrices​​ from sequencing data [sequence​​​‌ analysis].Bioinformatics41​3March 2025,​‌ 1-5HALDOIback​​ to text

International peer-reviewed​​​‌ conferences

• 22 inproceedingsA.​Amanda Carroll, L.​‌Leanne Mortimer, H.​​Hiren Ghosh, S.​​​‌Sandra Reuter, H.​Hajo Grundmann, K.​‌Karel Brinda, W.​​William Hanage, A.​​​‌Angel Li, A.​Andrew Purssell, B.​‌Bryan Coburn, A.​​Ashley Rooney, S.​​​‌Shola Able-Thomas, M.​Martin Antonio, D.​‌Derek Macfadden and A.​​Allison Mcgeer. P-397.​​​‌ Rapid Prediction of Genetic​ Relatedness of Escherichia coli​‌ Direct from Critical Care​​ Samples Using Metagenomic Sequencing​​​‌ Paired with Neighbour Typing​.Open Forum Infectious​‌ DiseasesIDWeek 202412​​IDWeek 2024 AbstractsSupplement_1​​​‌Los Angeles, CA, United​ StatesOxford University Press​‌2025, S357HAL​​DOIback to text​​​‌
• 23 inproceedingsM.Marie​ Pegard, A.Adela​‌ Poublan-Couzardot, P.Philippe​​ Barre, S.Simon​​​‌ de Givry, C.​Christine Gaspin, F.​‌Fabrice Legeai, F.​​Frédéric Choulet, C.​​​‌Christophe Klopp and B.​Bernadette Julier. Enhanced​‌ Genome Assemblies of French-Bred​​ Dactylis glomerata and Medicago​​​‌ sativa: Achieving High-Quality Tetraploid​ Genomes.Breeding and​‌ Genetic Improvement for a​​ Net-Zero FutureFCAG 2025​​​‌ - 36th Meeting of​ the EUCARPIA Fodder Crops​‌ and Amenity Grasses Section​​Aberystwyth, United KingdomSeptember​​​‌ 2025, 1-2HAL​back to text
• 24​‌ inproceedingsO.Ondřej Sladký​​, P.Pavel Veselý​​​‌ and K.Karel Břinda​. Towards Efficient k-Mer​‌ Set Operations via Function-Assigned​​ Masked Superstrings.Proceedings​​​‌ of the Prague Stringology​ Conference 2025PSC 2025​‌ - Prague Stringology Conference​​Prague, Czech RepublicAugust​​​‌ 2025, 26-40HAL​DOIback to text​‌

Conferences without proceedings

• 25​​ inproceedingsL.Léo Ackermann​​​‌, P.Pierre Peterlongo​ and K.Karel Břinda​‌. Towards space-efficient data​​ structures for large genome-distance​​​‌ matrices with quick retrieval​.DSB 2025 -​‌ Workshop Data Structures in​​ BioinformaticsPisa, Italy2025​​​‌HALback to text​back to text
• 26​‌ inproceedingsC.Charly Airault​​, C.Charles Deltel​​​‌, F.Florestan de​ Moor, E.Erwan​‌ Drezen, M.Meven​​ Mognol and D.Dominique​​​‌ Lavenier. Protein database​ search using Processing-in-Memory architecture​‌.IPDPS 2025 -​​ IEEE International Parallel and​​​‌ Distributed Processing SymposiumMilano,​ ItalyIEEE2025,​‌ 939-948HALDOIback​​ to text
• 27 inproceedings​​L.Lune Angevin,​​​‌ J.Josipa Lipovac,‌ R.Riccardo Vicedomini,‌​‌ P.Pierre Peterlongo and​​ E.Emeline Roux.​​​‌ Identification of bacterial strains‌ in gut microbiota: tools‌​‌ comparison and application.​​Journées du PEPI IBIS​​​‌ 2025Rennes, France2025‌, 1-18HALback‌​‌ to text
• 28 inproceedings​​L.Lune Angevin,​​​‌ P.Pierre Peterlongo,‌ R.Riccardo Vicedomini,‌​‌ A.Anne Siegel,​​ J.Jeanne Got and​​​‌ E.Emeline Roux.‌ Metagenomic taxonomic assignment using‌​‌ Nanopore reads, reconstruction of​​ metabolic networks and prediction​​​‌ of metabolite production.‌2025 - Journées Nutrition‌​‌ et Ecosystèmes MicrobiensRennes,​​ France2025HALback​​​‌ to text
• 29 inproceedings‌ F.Fantine Benoit,‌​‌ M.Mélody Temperville,​​ M.Manon Le Goff​​​‌, C.Claire Lemaitre‌, S.Sylvain Glémin‌​‌ and C.Claire Mérot​​. Detecting chromosomal inversions​​​‌ for population genomics: what‌ could be the optimal‌​‌ approach? GRC 2025 -​​ Gordon Research Conference on​​​‌ Ecological and Evolutionary Genomics‌ Lucques (Barga), Italy 2025‌​‌ HAL back to text​​
• 30 inproceedings F.Fantine​​​‌ Benoit, M.Mélody‌ Temperville, M.Manon‌​‌ Le Goff, C.​​Claire Lemaitre, S.​​​‌Sylvain Glémin and C.‌Claire Mérot. Detecting‌​‌ chromosomal inversions for population​​ genomics: what could be​​​‌ the optimal approach? Alphy‌ & AIEM 2025 -‌​‌ ALignement et PHYlogénie -​​ Approche Interdisciplinaire de l'Evolution​​​‌ Moléculaire Lyon, France 2025‌ HAL back to text‌​‌
• 31 inproceedingsF.Francesca​​ Brunetti and K.Karel​​​‌ Brinda. Optimized K-mer‌ Matching For Million-Genome Collections‌​‌ On Laptops.DSB​​ 2025 - Workshop Data​​​‌ Structures in BioinformaticsPisa,‌ ItalyMarch 2025HAL‌​‌back to textback​​ to text
• 32 inproceedings​​​‌F.Francesca Brunetti and‌ K.Karel Brinda.‌​‌ Rapid searches across million-genome​​ bacterial collections on laptops:​​​‌ a practical methodology for‌ point-of-care applications using k-mers‌​‌.SIM 2025 -​​ 53rd National Congress of​​​‌ the Italian Society of‌ MicrobiologyCatania, Italy2025‌​‌HALback to text​​back to text
• 33​​​‌ inproceedingsN.Nicolas Maurice‌, C.Claire Lemaitre‌​‌, C.Clémence Frioux​​ and R.Riccardo Vicedomini​​​‌. Investigating pre-assembly clustering‌ of HiFi reads for‌​‌ de novo assembly of​​ complex metagenomes.SeqBIM​​​‌Nantes, FranceNovember 2025‌HALback to text‌​‌
• 34 inproceedingsM.Meven​​ Mognol, F.Florestan​​​‌ de Moor, E.‌Erwan Drezen, Y.‌​‌Yann Falevoz and D.​​Dominique Lavenier. Evaluating​​​‌ Energy Efficiency of Genomics‌ Algorithms on Processing-in-Memory Architectures‌​‌.PECS 2025 -​​ International Workshop on Performance​​​‌ and Energy Efficiency in‌ Concurrent and Distributed Systems‌​‌Dresden, GermanyAugust 2025​​, 1-12HALback​​​‌ to text
• 35 inproceedings‌A.Alix Regnier and‌​‌ P.Pierre Peterlongo.​​ k-mer matrix compression.​​​‌DSB 2025 Pisa -‌ Workshop Data Structures in‌​‌ BioinformaticsPise, Italy2025​​HALback to text​​​‌
• 36 inproceedingsO.Ondřej‌ Sladký, P.Pavel‌​‌ Veselý and K.Karel​​ Brinda. Masked superstrings​​​‌ as a compact, indexable,‌ and dynamic representation of‌​‌ unconstrained k-mer sets.​​DSB 2025 - Workshop​​​‌ Data Structures in Bioinformatics‌Pisa, Italy2025HAL‌​‌back to text
• 37​​​‌ inproceedingsM.Mélody Temperville​, F.Fantine Benoit​‌, C.Claire Mérot​​, F.Fabrice Legeai​​​‌ and C.Claire Lemaitre​. SVJedi-Tag : a​‌ novel method for genotyping​​ large inversions with linked-read​​​‌ data.JOBIM 2025​ - Journées Ouvertes Biologie,​‌ Informatique et MathématiquesBordeaux,​​ France2025, 1-9​​​‌HALback to text​
• 38 inproceedingsT. K.​‌Tam Khac Minh Truong​​, D.Dominique Lavenier​​​‌, P.Pierre Peterlongo​ and K.Karel Břinda​‌. Optimized phylogenetic batching​​ of million-genome collections for​​​‌ reduced storage requirements and​ faster data retrieval.​‌SeqBIM 2025Nantes, France​​2025HALback to​​​‌ textback to text​
• 39 inproceedingsR.Riccardo​‌ Vicedomini and R.Rayan​​ Chikhi. Improved strain-level​​​‌ metagenome assembly for modern​ long reads.SeqBIM​‌ 2025 - Journées sur​​ les Séquences en Bioinformatique,​​​‌ Informatique et MathématiquesNantes,​ France2025HALback​‌ to text

Edition (books,​​ proceedings, special issue of​​​‌ a journal)

• 40 proceedings​SeqBIM 2025 Abstracts and​‌ Proceedings.SeqBIM 2025​​Nantes, France2025,​​​‌ 1-30HALback to​ text

Doctoral dissertations and​‌ habilitation theses

• 41 thesis​​M.Meven Mognol.​​​‌ Acceleration of bioinformatics algorithms​ on a Processing-in-Memory architecture​‌.Université de Rennes​​July 2025HALback​​​‌ to text

Reports &​ preprints

• 42 miscC.​‌Carole Belliardo, N.​​Nicolas Maurice, A.​​​‌Arthur Pere, S.​Samuel Mondy, A.​‌Alain Franc, M.​​Marc Bailly-Bechet, C.​​​‌Claire Lemaitre, R.​Riccardo Vicedomini, J.-M.​‌Jean-Marc Frigerio, F.​​Franck Salin, É.​​​‌Élodie Belmonte, D.​ J.David James Sherman​‌, P.Pierre Abad​​, C.Clémence Frioux​​​‌ and É. G.Étienne​ G.J. Danchin. Accurate​‌ MAG reconstruction from complex​​ soil microbiome through combined​​​‌ short- and HiFi long-reads​ metagenomics.September 2025​‌HALDOIback to​​ textback to text​​​‌
• 43 miscS.Stéphanie​ Bocs, C.Camille​‌ Carrette, J.Johann​​ Confais, S.Siegfried​​​‌ Dubois, L.Ludovic​ Duvaux, C.Christophe​‌ Klopp, N.Nicolas​​ Lapalu, P.Pauline​​​‌ Lasserre-Zuber, F.Fabrice​ Legeai, C.Claire​‌ Lemaitre, B.Benjamin​​ Linard, N.Nina​​​‌ Marthe, B.Baptiste​ Pierre, G.Gautier​‌ Sarah, F.François​​ Sabot, C.Christine​​​‌ Tranchant-Dubreuil and M.Matthias​ Zytnicki. A roadmap​‌ for the adoption of​​ pangenomics in agronomy.​​​‌2025HALDOIback​ to text
• 44 misc​‌F.Francesca Brunetti and​​ K.Karel Břinda.​​​‌ Optimized k-mer search across​ millions of bacterial genomes​‌ on laptops.November​​ 2025HALDOIback​​​‌ to textback to​ text
• 45 miscM.​‌Mabel Budia-Silva, A.​​Amanda Carroll, H.​​​‌Hiren Ghosh, A.​Allison Mcgeer, T.​‌Tommaso Giani, G.​​ M.Gian Maria Rossolini​​​‌, K.Karel Brinda​, W.William Hanage​‌, H.Hajo Grundmann​​, D.Derek Macfadden​​​‌ and S.Sandra Reuter​. Neighbour Typing Using​‌ Long Read Sequencing Provides​​ Rapid Prediction of Sequence​​​‌ Type and Antimicrobial Susceptibility​ of Klebsiella pneumoniae.​‌September 2025HALDOI​​back to text
• 46​​ miscR.Rayan Chikhi​​​‌, T.Téo Lemane‌, R.Raphaël Loll-Krippleber‌​‌, M.Mercè Montoliu-Nerin​​, B.Brice Raffestin​​​‌, A. P.Antonio‌ Pedro Camargo, C.‌​‌ J.Carson J Miller​​, M. B.Mateus​​​‌ Bernabe Fiamenghi, D.‌ P.Daniel Paiva Agustinho‌​‌, S.Sina Majidian​​, G.Greg Autric​​​‌, M.Maxime Hugues‌, J.Junkyoung Lee‌​‌, R.Roland Faure​​, K. D.Kristen​​​‌ D Curry, J.‌ A.Jorge A Moura‌​‌ de Sousa, E.​​ P.Eduardo P C​​​‌ Rocha, D.David‌ Koslicki, P.Paul‌​‌ Medvedev, P.Purav​​ Gupta, J.Jessica​​​‌ Shen, A.Alejandro‌ Morales-Tapia, K.Kate‌​‌ Sihuta, P. J.​​Peter J Roy,​​​‌ G. W.Grant W‌ Brown, R. C.‌​‌Robert C Edgar,​​ A.Anton Korobeynikov,​​​‌ M.Martin Steinegger,‌ C. A.Caleb A‌​‌ Lareau, P.Pierre​​ Peterlongo and A.Artem​​​‌ Babaian. Logan: Planetary-Scale‌ Genome Assembly Surveys Life’s‌​‌ Diversity.July 2025​​HALDOIback to​​​‌ text
• 47 miscA.‌Arya Kaul, F.‌​‌ W.Fernando W Rossine​​, K.Karel Brinda​​​‌ and M.Michael Baym‌. Novel genes arise‌​‌ from genomic deletions across​​ the bacterial tree of​​​‌ life.January 2026‌HALDOIback to‌​‌ text
• 48 miscV.​​Victor Levallois, Y.​​​‌Yoshihiro Shibuya, B.‌Bertrand Le Gal,‌​‌ R.Rob Patro,​​ P.Pierre Peterlongo and​​​‌ G. E.Giulio Ermanno‌ Pibiri. Kaminari: a‌​‌ resource-frugal index for approximate​​ colored k -mer queries​​​‌.May 2025HAL‌DOIback to text‌​‌
• 49 miscJ.Josipa​​ Lipovac, M.Mile​​​‌ Šikić, R.Riccardo‌ Vicedomini and K.Krešimir‌​‌ Križanović. MADRe: Strain-Level​​ Metagenomic Classification Through Assembly-Driven​​​‌ Database Reduction.May‌ 2025HALDOIback‌​‌ to textback to​​ text
• 50 miscS.​​​‌Sandra Romain, S.‌Siegfried Dubois, F.‌​‌Fabrice Legeai and C.​​Claire Lemaitre. Investigating​​​‌ the topological motifs of‌ inversions in pangenome graphs‌​‌.March 2025HAL​​DOIback to text​​​‌

Other scientific publications

• 51‌ inproceedingsL.Lune Angevin‌​‌, J.Josipa Lipovac​​, R.Riccardo Vicedomini​​​‌, P.Pierre Peterlongo‌ and E.Emeline Roux‌​‌. From bacterial taxonomic​​ classification at strain level​​​‌ to metabolic networks producibility‌ in gut microbiota.‌​‌2025 - 14th International​​ Gut Microbiology SymposiumClermont-Ferrand,​​​‌ France2025HALback‌ to text
• 52 inproceedings‌​‌C.Carole Belliardo,​​ A.Arthur Péré,​​​‌ A.Alain Franc,‌ J.-M.Jean-Marc Frigerio,‌​‌ N.Nicolas Maurice,​​ C.Clémence Frioux,​​​‌ C.Claire Lemaitre,‌ S.Samuel Mondy,‌​‌ M.Marc Bailly-Bechet,​​ P.Pierre Abad,​​​‌ D. J.David James‌ Sherman and E. G.‌​‌Etienne G.J. Danchin.​​ Enhancing Microbial Genome Reconstruction​​​‌ in Complex Environments by‌ combining Short-and Long-read Sequencing‌​‌.2025 - Journées​​ scientifiques Agroécologie et Numérique​​​‌Dijon, FranceJanuary 2025‌, 1-1HALback‌​‌ to text
• 53 inproceedings​​F.Fantine Benoit,​​​‌ M.Mélody Temperville,‌ C.Claire Lemaitre and‌​‌ C.Claire Mérot.​​​‌ Evaluating sequencing techniques for​ chromosomal inversion detection in​‌ non-model species: insights from​​ Coelopa frigida genomes..​​​‌GRS 2025 - Ecological​ and Evolutionary Genomics): Elucidating​‌ the Evolutionary Dynamics of​​ Adaptation in Fluctuating Environments​​​‌Lucques, Italy2025HAL​back to text
• 54​‌ inproceedingsT.Thomas Brazier​​, C.Claire Lemaitre​​​‌ and C.Claire Mérot​. Diversity of genomic​‌ structural variation across the​​ Tree of Life.​​​‌Jacques Monod Speciation: “A​ multidimensional view of speciation:​‌ bridging micro and macro-evolution”​​Roscoff, FranceOctober 2025​​​‌HALback to text​
• 55 inproceedingsT.Thomas​‌ Brazier, C.Claire​​ Lemaitre and C.Claire​​​‌ Mérot. Structural genetic​ diversity across the Tree​‌ of Life: development of​​ a long-read based pipeline​​​‌ for robust detection of​ SVs.PopGroup58Sheffield,​‌ United KingdomJanuary 2025​​HALback to text​​​‌
• 56 inproceedingsT.Thomas​ Brazier, C.Claire​‌ Lemaitre and C.Claire​​ Mérot. Structural genetic​​​‌ diversity across the Tree​ of Life: development of​‌ a long-read based pipeline​​ for robust detection of​​​‌ SVs.ALPHYLyon,​ FranceFebruary 2025HAL​‌back to text
• 57​​ inproceedingsM.Mabel Budia-Silva​​​‌, A.Amanda C.​ Carroll, H.Hiren​‌ Ghosh, A.Allison​​ Mcgeer, T.Tommaso​​​‌ Giani, G.Gian​ Maria Rossolini, K.​‌Karel Brinda, W.​​William P. Hanage,​​​‌ H.Hajo Grundmann,​ D.Derek R. Macfadden​‌ and S.Sandra Reuter​​. Neighbour Typing Using​​​‌ Long Read Sequencing Provides​ Rapid Prediction of Sequence​‌ Type and Phenotype in​​ Klebsiella pneumoniae sensu lato​​​‌ from Europe.Applied​ Bioinformatics & Public Health​‌ MicrobiologyHinxton, Cambridge, United​​ KingdomMay 2025HAL​​​‌back to text
• 58​ inproceedingsN.Nicolas Maurice​‌, C.Clémence Frioux​​, C.Claire Lemaitre​​​‌ and R.Riccardo Vicedomini​. Mapler: Assessing assembly​‌ quality in taxonomically-rich metagenomes​​ sequenced with HiFi reads​​​‌.2025 - Journées​ scientifiques Agroécologie et Numérique​‌Dijon, France2025,​​ 1-1HALback to​​​‌ text
• 59 inproceedingsO.​Ondřej Sladký, P.​‌Pavel Veselý and K.​​Karel Břinda. Masked​​​‌ Superstrings as a compact,​ indexable, and dynamic representation​‌ of unconstrained k-mer sets​​.RECOMB 2025 -​​​‌ 29th International Conference of​ Research in Computational Molecular​‌ BiologySeoul, Korea, South​​ Korea2025, 1-1​​​‌HALback to text​

Scientific popularization

• 60 inbook​‌D.Dominique Lavenier and​​ M.Marc Antonini.​​​‌ Stockage d'information sur ADN​.Le calcul à​‌ découvertCNRS EditionsJanuary​​ 2025HALback to​​​‌ text
• 61 bookD.​Dominique Lavenier. Stocker​‌ nos données sur ADN​​.Espace de sciences,​​​‌ RennesMarch 2025,​ https://www.editions-apogee.com/espace-des-sciences/719-stocker-nos-dones-sur-adn.htmlHALback to​‌ text
• 62 articleE.​​Emeline Roux and G.​​​‌Gaëlle Boudry. Biodiversité​ alimentaire, microbiote et bien-être​‌ : la recherche explore​​ les liens potentiels.​​​‌The Conversation FranceNovember​ 2025HALDOIback​‌ to text