Dr Liam J. McGuffin, Associate Professor in Bioinformatics
Biomedical Sciences Division, School of Biological Sciences, University of Reading UK
Friday, September 4th 2015, 2pm, LIRMM Bat 5 - Salle 1/124 (see map)
Predicting 3D models of proteins allows us to bridge the sequence-structure gap and cope with the sequence data deluge resulting from next-generation sequencing. Computational methods for predicting protein structures from sequence are relatively fast and inexpensive and they allow us to study proteins that are problematic to resolve experimentally, for example, those with long regions of intrinsic disorder or those with membrane spanning regions. For the majority of sequences, accurate 3D models of proteins can be generated, which may then be used to inform and direct experimental work. Predicted protein structures may allow you to infer gene function and can help you to understand more about the mutations associated with disease. We have developed an integrated suite of structural bioinformatics tools, IntFOLD, for protein fold recognition, 3D model quality assessment, intrinsic disorder prediction, domain prediction and ligand binding site prediction. Recently, we have also been working on the development of tools for 3D modelling of protein-protein interactions from sequences. Over the past 5 years we have applied our bioinformatics tools to investigate the molecular mechanisms of various human diseases as well as plant-pathogen interactions. I will be highlighting a few disease case studies and explaining how the effects of mutations can be studied through modelling of protein 3D structures and their interactions.