Section: New Results

Population structure and dynamics

We recently started a collaboration with the Pasteur Institute in Cambodia (Dr. P. Buchy) and the CIRAD at Montpellier (Dr. R. Frutos) on viral population structure and dynamics. In this context, we developed an exploratory statistical approach to characterise mutational patterns in viral populations. The basic idea is to use Multiple Correspondence Analysis (MCoA) on a multiple alignment of nucleic sequences. To this purpose, the alignment is encoded as a boolean table where rows correspond to the sequences and columns to the presence/absence of characters. This can be done simply by considering each of the four possible bases as a different character, or by considering only two possible states: one for the major base and one for all other (minor) bases. This technique turned out to be very effective in representing co-mutation patterns within a population of sequences. As for simple CoA, the plot is quite easy to interpret by biologists, both in terms of proximities of sequences and characters (i.e. mutations). Moreover, it has some strong relationships with parsimony phylogeny that need to be clarified. In [10] , we applied this technique to study the structure and time evolution of the Dengue virus (serotype I) population in Cambodia, using heterochronous sequence samples. Beside its methodological aspect, this work also introduced new topics related to population genetics in BAMBOO (e.g. the use of coalescent theory to reconstruct population dynamics). Another approach, widely used in ecology, to measure biodiversity and to determine the species composition of environmental samples is the "DNA-barcoding" technique. In collaboration with the Laboratoire d'Ecologie Alpine (Univ. Joseph Fourier Grenoble), we developed a new software for identifying new barcode markers and their associated PCR primers [19] .