ContactLib

The need to obtain structures similar to the new protein from these databases motivates the neighbor protein structure retrieval problem: given a query protein structure and a database of protein structures, retrieve all the structures in the database that are similar to the query structure. We introduce ContactLib, a complete contact group library, which can be used as fingerprints of protein structures to retrieve neighbor protein structures efficiently and effectively.

Download

  • README.txt: the readme document

  • contactlib-bin.Dec-7-2013.tgz: the ContactLib executables compiled for Linux-64bit

  • contactlib-scop40-2.03.Dec-7-2013.tgz: the pre-built ContactLib for SCOP40

  • contactlib-cullpdb40-res2.0-R0.25.Dec-8-2013.tgz: the pre-built ContactLib for CULLPDB40-RES2.0-R0.25

    • User Manual

  • To search similar protein structures from a ContactLib, call "./bin/runContactLib.sh IN_PDB_FILE IN_CONTACTLIB_DIR OUT_SCORE_FILE", where "IN_PDB_FILE" is the input PDB file containing the query protein structure, "IN_CONTACTLIB_DIR" is the input ContactLib directory containing the pre-built ContactLib, and "OUT_SCORE_FILE" is the output file containing the similarities scores. The pre-built ContactLib for SCOP40 is provided.

  • To build your own ContactLib from PDB files, call "./bin/buildContactLib.sh IN_PDB_DIR OUT_CONTACTLIB_DIR", where "IN_PDB_DIR" is the input directory containing all PDB files, and "OUT_CONTACTLIB_DIR" is the output directory containing the ContactLib.

    • Reference

  • Xuefeng Cui, Shuai Cheng Li, Lin He, Ming Li, Fingerprinting protein structures effectively and efficiently, Bioinformatics (2013)