D-I-TASSER (Distance-guided Iterative Threading ASSEmbly Refinement) is a new method extended from I-TASSER for high-accuracy protein structure and function predictions. Starting from a query sequence, D-I-TASSER first generates inter-residue contact maps, distance maps and hydrogen-bond (HB) networks using multiple deep neural-network predictors, including AttentionPotential (self-attention network built on MSA transformer) and DeepPotential. Meanwhile, it identifies structural templates from the PDB by the meta-threading LOMETS3 approach. The full-length atomic models are finally assembled by iterative fragment assembly Monte Carlo simultions under the guidance of I-TASSER force field and deep-learning contact/distance/HB restraints, where biological functions of the query protein are derived from the structure models by COFACTOR. The large-scale benchmark tests showed that D-I-TASSER generates significantly more accurate models than I-TASSER, especially for the sequences that do not have homologous templates in the PDB. D-I-TASSER server provides an optional D-I-TASSER-AF2 pipeline, which incorporates AlphaFold2 restraints with D-I-TASSER and generates models with average accuracy higher than both D-I-TASSER and AlphaFold2. The output model of D-I-TASSER server is given by both PDB format and ModelCIF format now. Please report problems and questions at our Discussion Board.
yangzhanglabumich.edu | (734) 647-1549 | 100 Washtenaw Avenue, Ann Arbor, MI 48109-2218