Protein fold predictions using pDomTHREADER based on the CATH domain library
pDomTHREADER is part of the latest implementation of the GenTHREADER tool suite available from the UCL Bioinformatics group. The GenTHREADER tools predict protein fold from sequence by combining profile-profile alignments, secondary structure gap penalties and both classic pair and solvation potentials using an optimised regression SVM model. Combining these features significantly improves the detection of useful template structures and the accuracy of sequence-structure alignments relative to other available methods. The pDomTHREADER implementation is specifically trained to discriminate structural superfamilies from within protein sequences.
Group Leader:David Jones