Protein modelling

From Opengenome.net
Revision as of 09:35, 26 August 2005 by 211.211.234.134 (talk)

http://biome.ngic.re.kr/ProteinModelling/


Problem (문제정의)

  • Build 3D structural models of given or interested proteins
    • Initial input: amino acid sequence of the target protein
    • Final output: its predicted 3D structure

Method

  • Find one or more structural templates of the target protein via sequence homology search against known structures (program: BLASTP or PSI-BLAST, database: PDB)
  • Download template structures in PDB format (web server: RCSB PDB)
  • Align the amino acid sequence of the target protein with that(those) of the template protein(s) (program: CLUSTALW, output: PIR format)
  • Build 3D structural models based on the multiple sequence alignment and the template 3D structure(s) (program: MODELLER, input: ATM, ALI, and TOP files)
    • manual build
      • manually convert file formats from PDB to ATM and PIR to ALI, respectively
      • write a TOP script file
      • run MODELLER
    • automatic build (some PDB and PIR files are not successfully converted to correct ATM and ALI files due to amino-acid sequence mismatch)
      • perl script to generate ATM, ALI, and TOP files from PDB and PIR files: txt, gz
      • command-line arguments in order: base_directory_path pir_file_name target_protein_id(exactly 4 letters) one_or_more_PDB_file_names(exactly 4-letter prefix)
      • run the perl script (e.g. ../bin/makeModellerInput.pl /home/user/Protein3DModelling/KCIP KCIP.pir KCIP 1QJA.pdb)
      • run MODELLER (e.g. mod7v7 KCIP.top)
    • example
      • ATM file: 5fd1.atm
      • ALI file: alignment.ali
      • TOP file: model-default.top
      • command: mod7v7 model-default.top
      • output PDB file (predicted model): 1fdx.B99990001.pdb
    • MODELLER manual
      • PDF, HTML
  • Display and compare the 3D structures of the model and the template proteins (program: RasMol)