Wrap-and-Pack: A New Paradigm for Beta Structural Motif Recognition with Application to Recognizing Beta Trefoils

Abstract

A method is presented that uses β-strand interactions at both the sequence and the atomic level, to predict beta-structural motifs of protein sequences. A program called Wrap-and- Pack implements this method and is shown to recognize β-trefoils, an important class of globular β-structures, in the Protein Data Bank with 92% specificity and 92.3% sensitivity in cross-validation. It is demonstrated that Wrap-and-Pack learns each of the ten known SCOP β-trefoil families, when trained primarily on β-structures that are not β-trefoils, together with three-dimensional structures of known β-trefoils from outside the family. Wrap-and-Pack also predicts many proteins of unknown structure to be β-trefoils. The computational method used here may generalize to other β-structures for which strand topology and profiles of residue accessibility are well conserved.