6 novembre 2020 - 13h30

Pedro Beltrao, European Bioinformatics Institute (EMBL-EBI), Hinxton

Résumé : Understanding how genetic variation contributes to phenotypic differences is a central question in biology. This requires models that can describe how rare and common genetic variability impacts on different cellular components and how these translate to trait variability. We measured growth phenotypes for ~1000 strains of E. coli and S. cerevisiae and to study the impact of rare variants we developed a set of variant effect predictors for protein stability, interaction affinity, and gene expression regulation. Using these tools we have determined gene and complex level burden scores and shown how the disruption of specific genes or complexes relates to trait variation. For human we have started to analyse the wealth of genome wide association studies (GWAS) that has grown to over 1000 human traits aggregated in the GWAS Catalog. Based on the principle that genes associated with the same phenotype tend to be involved in the same cellular processes, network-based methods have been used to expand and prioritize trait associated genes. I will present and benchmark a strategy able to identify groups of highly connected genes relevant for a phenotype of interest, starting from GWAS candidate genes. The network based gene-trait association scores allows for trait-trait similarity analysis and can be further combined with patient specific mRNA/protein data for further prioritization. I will focus on neurodegeneration diseases to illustrate the identification of shared aetiology and further integrate ALS SOD1 mutant patient mRNA/protein data to identify novel astrocyte candidate ALS linked genes. As a future perspective I will briefly discuss the need for approaches that can jointly integrate common and rare genetic variation to improve our understanding of trait variation.

Lieu : Amphithéâtre Charpak, Campus Pierre et Marie Curie, LNPHE, rez-de-chaussée tour 22

---

Séminaire externe de l'IBPS