Uncovering hidden data structures from large scale experiments is a common problem across many areas of biology from the analysis of sequencing data to the recording of neuronal populations. Standard methods of dimensionality reduction and clustering based on abstract data representations are often insufficient to capture biological complexity and are difficult to validate with statistical confidence. We will discuss model-based approaches to data classification in three biological contexts: deregulation patterns in RNA sequencing data, analysis of neuronal subtypes and the detection of synchronous neuronal populations. In all these contexts Bayesian inference can be used to support data classification with rigorous statistical evidence.

Sperm-dependent parthenogenesis, also called pseudogamy, is a reproductive strategy in which all-female species use the sperm ofmales, usually from another species, to activate their oocytes. The sperm DNA does not participate to the development of the zygote, whichproduces only females. We discovered a novel and unique reproductive strategy in the pseudogamous nematode species "Mesorhabditis belari" , which produces its own males at low frequency. Here, females are produced asexually while males are produced sexually. Hence, the males never transmit their genes to their female offspring. We described the mechanisms that allow the production of sexual males. We also used game theory to explain the 9% sex ratio found in this auto-pseudogamous species. We have established a collection of auto-pseudogamous and sexual species of Mesorhabditis which will allow us to explore the origin and consequences of this partial loss of males and asexuality. 

Macromolecular interactions are central to most biological processes. Protein docking aims to predict the most likely structural binding modes of interacting protein partners. Understanding how binding partners coevolved can provide essential clues to improve the structural prediction of protein interfaces. In the past few years, our team has contributed to the improvement of protein docking methods by combining evolutionary information with more traditional approaches. I will present how we analyzed the way interface structure coevolved [1] and developed InterEvDock, a server for protein-protein docking designed to integrate evolutionary information in the docking process [2, 3]. InterEvDock uses one of the most successful interface sampling algorithms to date, which relies on fast Fourier transforms, followed by a consensus scoring scheme to discriminate correct from incorrect interfaces. InterEvDock was benchmarked on a large dataset of docking targets based on unbound homology models [4]. Finally, I will discuss how we successfully applied this pipeline to…

During cell division, eukaryote chromosomes transform between a somewhat open interphase and somewhat closed metaphase conformation. I will discuss our group's work on studies of the mechanics of both these states of the genome.  I aim to discuss the role (and underlying mechanism) of SMC complexes in the folding, compaction, and topological organization of metaphase chromosomes, and also to describe how epigenetic marks appear to play a key role in controlling nuclear mechanics and topology.

Immune recognition of infected and malignant cells requires presentation on their surface of antigens (i.e. short peptides) by human leukocyte antigen class I (HLA-I) proteins, which are coded by one of the most polymorphic alleles in the human genome. The identification of clinically relevant, tumour-specific neoantigens (mutated antigens) is currently a highly sought-after goal in designing novel cancer immunotherapeutic strategies. Algorithms aimed at predicting peptide presentation by HLA-I proteins are therefore valuable tools to accelerate the validation of putative neoantigens. To tackle this problem, we resort to a framework of inference from aminoacid sequences based on Restricted Boltzmann Machines, probabilistic graphical models characterized by a layer of ‘feature’ variables. This approach ensures efficient prediction of what antigens can be presented along with their HLA-I binding specificity; furthermore, it can be used to study in a model-guided way the effect of mutations on antigen presentation.

The origin of maize (Zea mays ssp. mays) in the US Southwest remains contentious, with conflicting archaeological data supporting either coastal or highland routes of diffusion of maize into the US.Furthermore, the genetics of adaptation to the new environmental and cultural context of the Southwest is largely uncharacterized. To address these issues, we compared nuclear DNA from 32 archaeological maizesamples spanning 6000 years of evolution to modern landraces from across Mexico. We found that the initial diffusion of maize into the Southwest at about 4000 years ago likely occurred along a highland route, followedby gene flow from a lowland coastal maize beginning at least 2000 years ago. Our population genetic analysis also enabled us to differentiate selection during domestication for adaptation to the novel climatic andcultural environment of the Southwest, identifying adaptation loci relevant to drought tolerance and sugar content. 

Neuroendocrine axes, composed by brain neurohormones, pituitary hormones and peripheral gland hormones, regulate major physiological functions and key-steps of life cycles in Vertebrates. We investigate how whole genome duplication events (WGD) that occurred at the basis of the vertebrate lineage (“1R and 2R” for 1rst and 2nd Rounds of WGD) and at the basis of the teleost lineage (“3R” for 3rd Round of WGD) have impacted the molecular and functional evolution of neuroendocrine axes and their roles in the control of life cycles. Model fish species investigated are chosen for their phylogenetical, ecological, socio-economical relevance and the complexity of their life cycle.