What's MAMM?

MAMMMulti Alignment
and Multi Modeling
stands for Multi Alignment and Multi Modeling. The convenient access to homology modeling of protein structure is the focus of this web service. Various types of template-search, sequence-alignment, and modeling methods are integrated and detailed options are hidden behind graphical front-end. User input is very simple: target amino acid sequence in FASTA format and checkbox-type choice of programs for template-searching, sequence-alignment, and modeling. Returned output are structures in PDB format, a set of protein-like scores, and, if accessible, domain information, and protein-protein interaction data.
A user can make total of 72 combinations of options with three template-searching programs, eight sequence-alignment programs, and three modeling methods. The template-searching programs used here are HHpred, PSI-BLAST, and CS-BLAST. The eight sequence-alignment programs are seven stand-alone programs (ClustalW2, DIALIGN2, MAFFT, MUSCLE, PROBCONS, T-Coffee in Fast mode, and T-Coffee in M-Coffee mode) and own code in each template-searching program. Modeller, Modeller-Restraint Simulated-Annealing or MR-SA, and Pseudo-Quadratic-Restraint Simulated-Annealing or PQR-SA were integrated as modeling programs.
The modeling method of MR-SA and PQR-SA are CHARMM-interfaced protocols that have been developed by the authors. Using an initial structure made from template proteins, a simulated annealing simulation gives protein models. MR-SA needs restraint sets from Modeller program while PQR-SA makes its own spline potential functions based on quadratic functions.
Test run using 108 CASP9 targets showed a negative correlation of prediction quality and success rate. MR-SA, a chimera of our own initial structure and restraining potentials from Modeller, gave best result. But its overall success rate is the lowest. PSI-BLAST finds best templates for backbone, but its success rate is around half. HHpred template-search and PQR-SA method have relatively high success rate, but their results, as a whole, are of moderate quality.

What's new update?

Search template structure

1. PSI-BLAST and CS-BLAST update PDB sequence and template database (2012-01-3)
2. HHpred update hhsearch's HMM and template database (2012-01-03)

Multi alignment method

1. T-coffee update version 8.03 --> 9.01 (2011-11-16)

Contact us

If you have any problem/question/comment, you can contact us (E-mail)