CEthreader (Contact Eigenvector-based threader) is a new threading algorithm that combines contact-map with profile-based alignments for protein fold recognition. Starting from a query sequence, CEthreader first creates ab initio contact prediction by coupling deep MSA and convolutional residual network training through ResPRE. The predicted 2D contact-map is then converted to a set of single-body Eigenvectors through the Eigen-decomposition technique, which are subsequently integrated with the profile and secondary structure scoring function to guide the dynamic programming alignment process. Large-scale benchmark shows that the combination of contact-map and profile alignments can help improve the accuracy of fold-recognition in particular for proteins that lack close-homology templates in the PDB. If you have questions and comments on CEthreader, please post them at the Service System Discussion Board. ( >>more about CEthreader ...)
From May. 10 to Jul. 30, 2020 (CASP14 season), the CEthreader server will be closed to external users. Jobs already submitted will not be affected. We apologize for any inconvenience this may cause.
yangzhanglabumich.edu | (734) 647-1549 | 100 Washtenaw Avenue, Ann Arbor, MI 48109-2218