Fig. 1: Flowcharts for D-I-TASSER protein structure prediction.
From: Deep-learning-based single-domain and multidomain protein structure prediction with D-I-TASSER
a, The D-I-TASSER pipeline consists of four steps of deep MSA generation, template detection by meta-threading server, deep-learning-based spatial restraint prediction and full-length model construction with iterative REMC fragment assembly simulations. b, The pipeline of the multidomain structural modeling module consisting of domain boundary identification, domain-level threading and MSA collections and interdomain feature assembly.
