Today it is well known that the noncoding genome provides the organisms with the raw material for novel RNAs and protein products, thus playing an important role in the emergence of genetic novelty. The wide use of transcriptomics has revealed a high pervasive transcription of presumed noncoding regions, and a fraction of the resulting RNAs have been shown to be translated by ribosome profiling experiments. If an important fraction of them is expected to be degraded right-away, some of them may be neutral or deleterious and, therefore, will be short-lived in evolutionary history, while a small fraction of them, if they provide an advantage to the organism, will be fixed and established as novel genes. These results attribute a central role to the noncoding genome in the emergence of new genes and the development of pathologies. However, the mechanisms governing these processes remain unclear, though they are essential to understand (i) the evolutionary forces governing the emergence of these new genes and the evolution of genomes and (ii) their role in some cancers, or other human pathologies. In this project, we aim at investigating with integrative approaches the potential of the noncoding genome to give rise to novel peptides or novel protein bricks in order to better understand its role in the evolution of organisms and the development of human diseases.

In this project, we aim at characterizing the competition occurring between functional and nonfunctional (or nonspecific) protein interactions in the cell. We work on the analysis of the physical and evolutionary properties (extracted from docking simulations and sequence analyses) of the interfaces formed either by functional or nonfunctional partners in order to better understand how selection operates on interaction energy landscapes to regulate the severe competition occurring between the proteins in the crowded cell. In such a crowded environment, a protein has to deal with numerous nonfunctional interactions and has evolved to bind the right partner in the right way (positive design) and to prevent misassemblies and interactions with nonfunctional partners (negative design). Positive and negative design, therefore, operate on the whole surface of proteins where, the binding sites but also the rest of the surface are designed so that nonfunctional interactions do not prevail over functional ones. To address these questions, we develop methods and representation frameworks that allow one to probe the interaction potential as well as the sequence and structural properties of protein surfaces.

Characterizing how environmental pressures have shaped the diversity of protein folds and functions is essential to understand the structural bases of molecular innovation and finally, biodiversity. Proteins are molecular Legos of protein bricks whose combination dictates the outcoming protein fold and function. Until recently, it was suggested that most folds have already been inventoried in public databases. Nevertheless, metagenomics-based analyses and genome sequencing of species living in extreme environments suggest that there are still new structural spaces to be discovered. These studies report important ratios of Orphan genes that cannot be related to any known protein fold. Notably, archaeal genomes usually contain about 30% of Orphans. Some of them have been shown to confer new traits and play significant roles in species adaptation to the environment. Are these Orphan made of already known protein Lego bricks or do they include new ones? Do Orphans fold into novel protein folds? How do environmental constraints shape the protein structural space? We aim to investigate the protein diversity of archaea from a wide variety of habitats, including extreme ones. The recent DeepMind's program AlphaFold2 offers an unprecedented opportunity to study the relationship between environmental constraints and the 3D repertoire of folds of a genome/environment. In this project, we address these questions by exploring a large dataset of thousands archaeal genomes and MAGS selected to cover the archaeal tree of life and ecosystem types, along with the annotation and study of their protein folds with AlphaFold2.

In this project, we aim to study the evolution of metabolic networks using comparative genomics combined with methods coming from structural bioinformatic methods, originally developed to detect co-evolving residues. We applied our method on 368 fungal genomes to identify clusters of co-evolving enzymes. This enabled the detection of an ancestral and connected network of enzyme activities on which were grafted small connected subnetworks of co-evolving (i.e. sharing similar phylogenetic profiles) enzyme activities, thus providing different metabolic capacities to a peculiar species. Interestingly, we showed that the ancestral enzymatic core covers almost all pathways found in fungi suggesting that precursors of these pathways were already existing in the ancestor (Andriamanga et al, in preparation).