Clustering of functional protein loop conformations: applications to kinases, antibodies, and Ras: a QBI online seminar with Roland Dunbrack

The QBI Online Seminar Series presents Roland Dunbrack, researcher at the Institute for Cancer Research at the Fox Chase Cancer Center. Dr. Dunbrack obtained his PhD from Harvard University under the direction of Martin Karplus and Jack Strominger followed by postdoctoral work with Fred Cohen and Ken Dill at UCSF. He developed an interest in structural bioinformatics as a graduate student, and his backbone-dependent rotamer library is widely used in protein structure prediction and design. He has continued to work in structural bioinformatics and protein structure prediction as a faculty member at the Fox Chase Cancer Center in Philadelphia where he has worked since completing his postdoc in 1997.

Abstract

Many proteins carry out their functional roles through loop regions between the elements of regular secondary structure. We have applied density-based clustering of protein loop conformations with a backbone dihedral angle metric and electron density cutoffs. Our clustering algorithm exploits unique properties of the Ramachandran map. Clustering of the activation loop of protein kinases leads to an 8-way classification system of active and inactive structures. We have used the clustering of antibody CDR loops to develop RosettaAntibodyDesign (RAbD) and redesigned a COV-1 spike protein antibody into several antibodies that bind COV-2 spike proteins at nanomolar affinities.