MoRFPred-plus: Computational Identification of MoRFs in Protein Sequences using Physicochemical Properties and HMM profiles

“…The EXP53 set was assembled by combining sequences from three different studies and includes sequences containing experimentally validated MoRFs that are disordered in isolation. These sets were previously used to develop predictors including OPAL, MoRFchibi‐web, MoRFpred‐plus, MoRFchibi, and MoRFpred . Altogether, there were 938 sequences in training and test sets.…”

“…To further improve the model performance, we combined MoRFpred‐plus and MoRFchibi with the proposed model, since they were constructed using complementary features and learning algorithms. MoRFpred‐plus targets the properties of disordered regions flanking the MoRF residue to identify MoRFs whereas MoRFchibi utilizes the similarity, composition, and contrast information of MoRFs and non‐MoRFs together with different SVM kernels to predict MoRFs. To calculate the scores for each residue, we applied the common averaging principle where all scores are added and divided by the number of models used (Figure S3, Supporting Information).…”