A Bi-Recursive Neural Network Architecture for the Prediction of Protein Coarse Contact Maps
CSB '02 Proceedings of the IEEE Computer Society Conference on Bioinformatics
Introduction to Bioinformatics
Introduction to Bioinformatics
Bioinformatics
Reconstruction of 3D Structures From Protein Contact Maps
IEEE/ACM Transactions on Computational Biology and Bioinformatics (TCBB)
Fault tolerance for large scale protein 3D reconstruction from contact maps
WABI'07 Proceedings of the 7th international conference on Algorithms in Bioinformatics
| Hi-index | 0.00 |
Prediction of intramolecular residue-to-residue contacts is a difficult problem. In this paper we consider a related subproblem: the fragment-to-fragment contact prediction problem. A fragment contact map is a coarse-grained approximation of a residue contact map. Although fragment contact maps are approximations, they still contain most information about residue contacts and, most importantly, they are more tolerant to errors than contact maps (at least random error). We compare with an extensive set of experimental tests the error tolerance of fragment contact maps with respect to the error tolerance of residue contact maps. Moreover, as a simple concrete application, we show how to build a fragment contact predictor from a residue contact predictor and we show that in some cases this technique can improve contact predictions.