MM-align: Aligning protein complex structures by dynamic programming

MM-align is an algorithm for structurally aligning multiple-chain protein-protein complexes. The algorithm is built on a heuristic iteration of a modified Needleman-Wunsch dynamic programming (DP) algorithm, with the alignment score specified by the inter-complex residue distances. The multiple chains in each complex are first joined, in every possible order, and then simultaneously aligned with cross-chain alignments prevented. The alignments of interface residues are enhanced by an interface-specific weighting factor. An optimal alignment between two complexes, as well as the overall TM-score, will be reported for each comparison.

What is the difference between TM-align and MM-align? TM-align is for aligning monomer protein structures while MM-align is designed for aligning multiple-chain protein complex structures. Although one can still use TM-align to align protein complexes after manually joining the chains, this will lead to suboptimal alignments with unphysical cross alignments. Therefore, the best result will be obtained if one uses TM-align to monomer structures and MM-align for multimer structures.

MM-align on-line

Format of input files:
Each complex should be in the standard PDB format with the two chains seperated by a "TER" (Click here for an example of the format).

Update history (please always download the newest version of MMalign.f and MMalign programs):

2009/09/30: fixed a bug in the output display.

Click MMalign to download the executable program for Linux system. Simply running the program itself will give you a brief instruction of how to use it. (In case that you are using a 32 bit computer and you get warning message of "bash: ./MMalign: cannot execute binary file", you need to download a 32-bit version of MMalign which is a little bit slower.)

Click MMalign.f to download the source of the entire program coded in Fortran77, which you can compile in your computer.

Click benchmark to download a list of 205 non-homologous proteins taken from PDB (sequence identity<30%) and a larger dimer dataset; these two sets of proteins was used as a benchmark test of MM-align method.

References:
S. Mukherjee, Y. Zhang, MM-align: a quick algorithm for aligning multiple-chain protein complex structures using iterative dynamic programming. Nucleic Acids Research 2009; 37: e83 (Download PDF file and supporting materials).