Research

My works may be grouped in three topics (see below). My past research was on population dynamics of small mammals and of birds, chaos in (meta)population models, and the evolution of dispersal (see below).

Phylogenetic Diversification

This has been my main research topic for twelve years. The main idea driving my research in this area is that phylogenetic trees give information about past events of speciation and extinction—the drivers of the diversity of Life. This approach is attractive because the fossil record, long considered as the best record of past diversification events, has many pitfalls, especially for terrestrial organisms. However, phylogenies of recent species must be analysed with care to infer speciation and extinction rates. I have developed several methods in order to combine phylogenetic and taxonomic data [27], to test the effects of species traits on speciation rate [32], or to fit a given theoretical distribution to the distribution of branching times [12].

(A) A tree simulated with speciation probability equal to 0.005 and extinction probability equal to 0.001. (B) The same clade after removing the extinct lineages (from [29]).


I have also devoted some effort to assess the statistical performance of these methods and others [29, 36]. Recently, I have developed equations for the general time-dependent birth–death model; from these I have derived algorithms to simulate phylogenetic trees under any model of diversification, and have developed a method for fitting these models to phylogenetic data. These new results allow to quantify the limits of inference on diversification with phylogenies.

Phylogenetic Analysis of Comparative Data

My contribution to this field was to develop the use of generalised estimating equations (GEEs) for the analysis of comparative data in collaboration with Julien Claude. GEEs can be seen as a generalisation of generalised least squares (GLS, a statistical procedure for fitting regression models to correlated observations) to non-normal responses (in the same way that generalised linear models generalise ordinary least squares).

Procedure for analysing comparative data in a phylogenetic framework using generalised estimating equations (from [24]).


Julien Claude also worked during his Ph.D. on the geometric assessment of morphological evolution in turtles [25, 31].

We applied the GEE-based comparative method to the issue of life history evolution in cichlid fishes in African great lakes [37]. This work showed evidence for parallel evolution towards decreased fecundity in the three lakes, Tanganyika, Malawi, and Victoria. The phylogenetic reconstruction further evidenced that this evolution occurred independently in pelagic and rocky habitats within a lake. Strikingly, other groups of pelagic fishes (including in the same lakes) display high fecundity with numerous small eggs. The fact that cichlids in the same habitat evolved in a different direction indicates the importance of lineage-effect and that natural selection does not act in a simple way.

Coalescence and Past Population Dynamics

Kingman's coalescent is a mathematical model of genealogical relationships in large populations. There are similarities between these genealogies and the phylogenies used to depict the relationships among species, and both can be studied with tree-like structures. Furthermore, DNA sequences are the main data used to infer them [21]. I started to work on the coalescent during a collaboration with Johan Michaux. Palearctic mammals have undergone drastic changes in distribution during ice ages. Johan has conducted a large scale investigation to track these changes with genetic data. I contributed some coalescent analyses on three species: Apodemus sylvaticus [26], A. flavicollis [30]), and Clethrionomys glareolus [34]. My current interest in this topic is in modelling complex population growth with DNA sequences.

Past Projects

Large-scale Population Dynamics of British Birds

This was my post-doc project, a collaboration between the University of East Anglia (Norwich) and the British Trust of Ornithology (BTO). The main contribution from this project was a large scale study on the patterns of dispersal in British birds using the ringing database of the BTO [10]. I also developed a method for fitting continuous dispersal models to distribution of dispersal distances ([22]; ask me for the R code), as well as large scale analyses of population synchrony [15, 20], breeding success [18], and density dependence [23].

Chaos in Metapopulation Dynamics

The issue of chaotic dynamics in natural populations attracted my attention in the context of spatial structure. I found out that most predictions of chaos are very sensitive to some assumptions of the models used for simulations [7, 8].

Population Dynamics of the Mediterranean Pine Vole

This was my Ph.D. work focused on the Mediterranean Pine Vole (Microtus duodecimcostatus). The interest of this species is its relatively stable population dynamics compared to others voles which have "cyclic" dynamics [1–6, 13].
Back to main