Current projects at XTBG

Since I arrived at XTBG in June 2012 as PI of the Ecological Evolution group, I have been focusing on the mechanisms of diversification of tropical taxa, and on the role of adaptation and hybridization in speciation using Ficus and Fagaceae as model species.
I also develop bioinformatic tools for the analysis of Next-Generation Sequencing data. (see here for available tools).
I am supervising several postdoc and student projects:

For more information on the Ecological Evolution group, see here

Research direction 1: The role of natural selection and local adaptation in speciation

Speciation and reinforcement in Neotropical spiral gingers.

I also work on Neotropical spiral gingers (genus Costus), a group of plants that went through one of the fastest radiations known after the colonization of the Neotropics. In particular, I investigated the two species C. pulverulentus and C. scaber that diverged recently and share a large area of sympatry. These species present isolation mechanisms compatible with a model of speciation with reinforcement, also suggesting a role of natural selection. The objective of this project is to test the reinforcement hypothesis using two approaches:

This work led by K.M. Kay at UCSC is still ongoing, but a genetic map containing 1275 RAD markers distributed over nine linkage groups (the number of chromosomes in Costus) has been built and the QTL mapping is under way.

Speciation, phylogeography and adaptive radiation of Caribbean lizards

Different ecotypes of Anolis roquet in Martinique
Anolis roquet Anolis roquet
Anolis roquet Anolis roquet

During my postdoctoral position at Bangor University with Prof Roger S. Thorpe, I have been investigating the relative role of geographic isolation (allopatry) and natural selection in population differentiation and speciation. The island of Martinique (in the French West Indies) and its endemic anole, Anolis roquet offer a great opportunity to disentangle the relative role of these two factors. Indeed, Martinique once was composed of five precursor islands that joined to form present day Martinique about two million years ago. Four distinct lineages of A. roquet evolved in allopatry on these precursor islands for up to 8 million years before coming into secondary contact when the islands joined. Furthermore Martinique is a mountainous island and is characterised by very sharp ecological gradients from dry coastal scrubland to wet mountain rainforest. Anoles living in these different habitats are subject to very different selective pressures and differ greatly morphologically. This system allows us to measure population differentiation across lineage and habitat contact zones and hence to look independently at the effect of these two factors.
Using mitochondrial DNA, microsatellites and morphology, we measured the population differentiation along independent transects across habitat and/or lineage contact zones. In all but one transects, we demonstrated that the geological history had no effect on the population differentiation, while habitat differences led to a strong reduction in gene flow. We concluded that contrarily to the expectations, this lizard had a better fit to the ecological speciation than to the allopatric speciation model, pointing to an important role of natural selection in speciation.

To investigate the generality of the patterns observed in Martinique, I am using the other species of the roquet group (A. luciae in St Lucia, A. griseus and A. trinitatis in St Vincent, A. aeneus and A. richardii in Grenada and A. bonairensis in Bonaire) as natural replicates. I am also interested in the phylogenetic relationships of the roquet group and for islands with two species of anoles (St Vincent and Grenada) to the comparative phylogeography of these species.

Finally I am also interested in phylogeny and adaptive radiation of the dwarf geckos Sphaerodactylus fantasticus and Sphaerodactylus vincenti in the Lesser Antilles.

Research direction 2: Tool development in evolutionary genetics of non-model organisms.

Evolutionary genetics work on non-model organisms has often been limited by the lack of genomic resources or of bioinformatic tools adapted to the analysis of species for which no reference genome exist.

For instance, the analysis of gene expression by RNA sequencing requires as a first step a de-novo assembly of the transcriptome, while a simple mapping of reads to a reference genome is sufficient for model organisms. De-novo transcriptome assembly of non model organisms present different challenges from genome assembly, in particular the non even coverage between different transcripts, while an even coverage is expected by genome assembly software. While working at the University of Geneva with Dr. Juam Montoya-Burgos, I developed two bioinformatic tools to optimize the assembly of de-novo transcriptome sequencing of non-model organisms from Next-Generation Sequencing (NGS) data. The first, called multiple-k method, relies on the fact that different k-mer values are needed to assemble transcripts with different coverage. The second, called Scaffolding using Translation Mapping (STM), makes use of a reference proteome to help assemble transcriptomes. Since sequence divergence at the amino-acid level is much lower than at the nucleotide level, it is possible to use the known proteome of distant model species to improve the transcriptome assembly of species with no reference.
The paper describing the methods and the source code of the pipeline are available here

Another problem generally encountered when working on non-model organisms is the lack of a sufficient number of variable markers for phylogenomics and population genomics. The development of NGS technology is changing this situation, and I am currently developing two alternative methods based on the Illumina MiSeq platform to generate a large amount of sequence information in two plant groups, the genus Ficus, and the Fagaceae. These methods will generate markers usable for both multilocus phylogenetic approaches and population genomics.

I also develop bioinformatic tools for the analysis of Next-Generation Sequencing data. For instance, because most tools available for RAD-seq analysis (or similar methods) don't allow the recovery of full haplotypes at each RAD locus, important information that could be used for phylogenetic analyses is lost. I am working on a tool to extract full haplotype information from NGS reads that generates input files for IMa2, a software making full use of the haplotype information.
Finally, I am working on a method to reconstruct phylogenetic relationships directly from raw NGS reads without preliminary assembly or alignment.

Research direction 3: Evolution of viviparity in Zootoca vivipara

My PhD work was part of a wider project aimed at understanding the evolution of reproductive modes in the common lizard Zootoca vivipara. Indeed, this lizard is a rare example of vertebrate species exhibiting a reproductive bimodality, i.e. where both oviparous and viviparous populations coexist within a single species (see figure below). This scarce situation offers a great opportunity to study this major evolutionary transition.

Zootoca vivipara eggs

A) Young of the viviparous form just after parturition.
B) Hatching of the oviparous form, about 40 days after egg-laying.
Photos by Benoît Heulin.

The first part of the project consisted to understand the geographical distribution of the various oviparous and viviparous populations of the European common lizard (Zootoca vivipara), to identify the distinct lineages existing in this species, and to elucidate the phylogenetic relationships of these lineages. Using mitochondrial DNA, I identified two oviparous and four viviparous lineages, and inferred that several shifts between reproductive modes occurred in this species. Furthermore, the most parsimonious hypothesis suggests that a reversal from viviparity back to oviparity occurred. However, more work using nuclear markers is needed to confirm this hypothesis, since preliminary analyses using AFLP suggested that introgression between the different lineages might have obscured the evolutionary history of this species.
In parallel, we conducted functional investigation to understand the differences between the oviparous and viviparous females. Both histological and endocrinological studies showed that very little differences existed between these two form, except for the eggshell characteristics, and the development of oviduct glands (at the origin of the eggshell) during vitellogenesis.
To understand these differences, I am currently conducting a comparative transcriptomic approach. The comparison of the transcriptional differences between oviducts of oviparous and viviparous females should shed light on the genetic mechanisms at the origin of the evolution of viviparity. This work is conducted in collaboration with Dr Benoît Heulin (UMR CNRS 6553).