Proteome Diversification in Evolution

Our research projects aim at decrypting protein sequences and structures to get insights into the fundamental mechanisms of life's machinery and optimally guide biological intervention. What are the sites in a protein crucial for selecting its cellular partners and binding to them? How can they be modified to modulate protein functioning? We address these questions through a multidisciplinary approach at the interface of genomics, transcriptomics, proteomics, evolution, artificial intelligence, and molecular modelling.  

Specifically, we focus on the protein diversity generated by the alternative initiation and termination of transcription and alternative splicing. These mechanisms produce insertions, deletions, and substitutions, sometimes substantial, that can modulate the shape, strength, stoichiometry, and specificity of protein interactions in multi-cellular eukaryotes. We study, mine, and learn from these variations to reconstruct the diversification of highly specialised proteomes in evolution. How did the proteomes responsible for complex behavioural traits like the ability of humans and zebra finches to learn "language" from their peers expand in evolution? Through the alternative usage of evolutionary ancient "themes" or protein innovation?

Our research expands through several peripheral and inter-connected scientific objectives. Ongoing and incubating projects include the prediction of protein and protein complex functional states, combining geometric manifold learning and natural language processing, for globular as well as non-globular proteins, the mapping of sequence variations to organismal-level phenotypes in the low data regime, with a focus on ageing, and the discovery of co-translational mechanisms from sequence signals.