Population Genetics, Developmental Transcriptomics, Bioinformatics
The evolution of breeding systems
The evolution of breeding systems is of great importance for the ecology and evolution of natural organisms. In plants, variation in mating systems is often associated with different features of the life history or demography of populations. Yet, reproductive strategies themselves can evolve within a single species and transitions of breeding systems are commonly observed at the phylogenetic level. In flowering plants, the most common transition is the one from outcrossing to self-fertilization and a rich body of empirical and theoretical literature has accumulated describing the selective processes driving these transitions. Shifts to pre-dominant self-fertilization have a profound impact on developmental, demographic, genomics, and ecological features of the biology of a species. However, no statistical method exists that allows inferring the age of shifts in mating systems using genome-wide polymorphism data. In this project we develop a model-based approach that aims at jointly estimating changes in rates of self-fertilization and fluctuations in population sizes. Precise dating of transitions to self-fertilization will allow us to better understand the climatic conditions in which these major evolutionary changes occur and to evaluate how long self-fertilizing species manage to maintain themselves in fluctuating environments.
Developmental transcriptomics: statistical analyses of time-lapse and single-cell transcriptome data
Temporal and spatial analysis of transcriptome data is a powerful approach to study how gene regulatory networks control organ development. Transcriptomics data obtained from time-lapse and single-cell experiments therefore represent an exciting opportunity to monitor early developmental pathways at the genetic level and the establishment of cellular identities. In our group, we develop statistical and computational tools for the analyses of time-series bulk-RNAseq and single-cell RNAseq (scRNAseq). Current work is focused on the identification of developmental stage and cell-type specific genes and on the comparative analyses of developmental transcriptome data between different species. We apply the methods developed in our lab to empirical data through a collaboration with the group of Prof. Miltos Tsiantis on the comparative analysis of leaf development between different plant species.
Reconstructing the demographic histories of natural populations using genetic variation
Population genetics theory provides theoretical models that establish a relation between genetic variation and major demographic features such as population sizes, population splits, and admixture events. Given empirical measures of genetic variation from natural populations, these theoretical results allow to calculate the likelihood of competing hypotheses regarding the evolutionary history of a species and to quantify important features of these histories such as the age of divergence or the intensity of gene flow between populations. These demographic reconstructions represent important ecological information about the history of natural populations but also provide valuable null hypotheses for studies aiming at identifying the effect of natural selection using genetic variation. In my group, we specialized in the reconstruction of demographic histories using large-scale genome-wide variation datasets. Besides developing pipelines and strategies to conduct population genomics analyses we are especially interested in understanding the impact of predominant self-fertilization and multiple merger on demographic inference. We apply these reconstruction approaches to Cardamine hirsuta, Drosophila melanogaster, and Copepods populations sampled in hydrothermal vents.
Expertise in Biostatistics and Bioinformatics
We provide help and support to the department and institute in issues related to biostatistics and bioinformatics. Our expertise in biostatistics covers design of experiment, statistical testing, data visualization, statistical modeling, and machine learning, as well as stochastic simulations. We also actively develop stable and reproducible bioinformatics pipeline (DNAseq, RNAseq, Chipseq).