HMMSTRTM: A hidden Markov model for local structure prediction in globular and membrane associated proteins
MotivationWe present HMMSTRTM, a Hidden Markov Model (HMM) that is useful for predicting topology of trans-membrane (TM) proteins. HMMSTRTM provides additional prediction categories of TM regions provided by the PDBTM corpus such as transmembrane beta sheets, coils, and reentrant loops. ResultsHMMSTRTM is competitive with existing TM protein topology predictors like TMHMM, it correctly predicts at least half the residues in 96.18% of all transmembrane helices in a cross validation dataset. AvailabilityModel architecture, source code, and supplementary figures are made available on github: github.com/TiburonB/HMMSTRTM. Contactbystrc@rpi.edu
bioinformatics↗