MRFy is Haskell software for protein remote homology prediction
You can read about our experience developing MRFy in the Haskell lazy functional programming language in our ICFP Experience Report.
You can access the source code for MRFy on Github.
If you use MRFy, please cite: N. Daniels, A. Gallant, N. Ramsey and L. Cowen, MRFy: Remote homology detection for beta-structural proteins using Markov random fields and stochastic search, Proceedings of ACM-BCB 2013.
MRFy relies on two 'helper' software packages: SSAnnotate, which is used to identify consensus beta-strands within a template structural alignment, and SMURFBuild, which is a fork of HMMER 3's hmmbuild needed to train a Markov random field for MRFy (or SMURF). MRFy also requires that a training alignment be produced using a structure aligner; we recommend Matt.