| November 18, 16:00
Evolution in the protein universe and epistasis in fitness landscapes
Our understanding of the structure of real genotype fitness landscapes at
different scales is still incomplete and requires new methods and models. We
present a model of long-term protein evolution that, once fitted to data from
prokaryotic protein alignments, shed light on the large-scale epistatic structure
of the protein universe. We also propose a natural measure of epistasis inspired
by this model, namely the correlation of fitness differences, and show how it can
be used both to measure local and global epistasis in experimental landscapes
and to discriminate between theoretical models. Finally we discuss how chains,
i.e. constrained paths in the landscape, are related to the epistasis and the
structure of the interactions.
Universite Pierre et Marie Curie and College de France
Biocentre, Ground Floor Lecture Hall
Contact: not specified