●I-TASSER ●QUARK ●LOMETS ●COACH ●COFACTOR ●MUSTER ●SEGMER ●FG-MD ●ModRefiner ●REMO ●SPRING ●COTH ●BSpred ●SVMSEQ ●ANGLOR ●BSP-SLIM ●SAXSTER ●ThreaDom ●ThreaDomEx ●EvoDesign ●GPCR-I-TASSER ●BindProf ●BindProfX ●ResQ ●IonCom ●STRUM

●TM-score ●TM-align ●MMalign ●NWalign ●EDTSurf ●MVP ●MVP-Fit ●SPICKER ●HAAD ●PSSpred ●3DRobot ●I-TASSER-MR ●NeBcon

●BioLiP ●E. coli ●GLASS ●GPCR-HGmod ●GPCR-RD ●GPCR-EXP ●TM-fold ●DECOYS ●POTENTIAL ●RW ●HPSF ●CASP7 ●CASP8 ●CASP9 ●CASP10 ●CASP11 ●CASP12

I-TASSER Decoy sets

• The I-TASSER Decoy Set I was taken from the trajectories of the I-TASSER simulations which include 12,500-32,000 raw decoys for each protein target. The backbone structure is built by I-TASSER ab initio simulation and the side-chain atoms are added using Pulchra. This decoy set was used to test the SPICKER clutering program for identifying best near-native structures.
  Please refer to the following paper for detail of the I-TASSER Decoy Set I:
  
  • S Wu, J Skolnick, Y Zhang: Ab initio modeling of small proteins by iterative TASSER simulations. BMC Biology 2007, 5: 17.
  
  

• The I-TASSER Decoy Set II is a structurally non-redundant set of protein structures which were selected from the Decoy Set I. Each protein target contains 300-500 decoys. Each decoy structure was first generated by I-TASSER Monte Carlo simulations and then refined by GROMACS4.0 MD simulation using OPLS-AA force field with the purpose of removing steric clashes and improving hydrogen-bonding network. This decoy set was used to test the ability of atomic potentials, including RW and RWplus potentials, for near-native structure identification and RMSD-energy correlation.
  Please refer to the following paper for detail of Decoy Set II:
  
  • J Zhang and Y Zhang, A Distance-Dependent Atomic Potential Derived from Random-Walk Ideal Chain Reference State for Protein Fold Selection and Structure Prediction. PLoS One, vol 5, e15386 (2010).

QUARK Decoy sets

• The QUARK Decoy Set was generated by the QUARK based ab initio folding simulations without using template structures. It contains 145 non-homologous proteins each with 5,000 decoys randomly selected from the simulation trajectories. The QUARK simulations were made on the backbone conformations, with the side-chain atoms added and further refined by the ModRefiner program.
  Please refer to the following paper for detail of this decoy set:
  
  • Dong Xu, Yang Zhang, Ab initio protein structure assembly using continuous structure fragments and optimized knowledge-based force field. Proteins, 2012, 80, 1715-1735 (download the PDF file and Support Information).
  • Dong Xu and Yang Zhang. Improving the Physical Realism and Structural Accuracy of Protein Models by a Two-step Atomic-level Energy Minimization. Biophysical Journal, vol 101, 2525-2534 (2011). (download the PDF file).

• The 3DRobot Decoy Set was generated by the 3DRobot, a program devoted for automated generation of diverse and well-packed protein structure decoys. The decoy set contains 200 non-homologous proteins each with 300 decoys that are evenly distributed in the RMSD space from 0 to 12 Anstrongs.
  Please refer to the following paper for detail of this decoy set:
  
  • Haiyou Deng, Ya Jia, Yang Zhang. 3DRobot: Automated Generation of Diverse and Well-packed Protein Structure Decoys. Boinformatics, in press (2015). (Download the PDF file and Support Information).

• The CASP8 Decoy Set contains the top 100 structural decoys generated by I-TASSER in CASP8, for all 121 protein domains that were finally assessed by the accessors. The decoys were ranked based on the structure density of the SPICKER clusters and 'model[1-5].pdb' are the structure models that were submitted to CASP8 by Zhang-Server.
  Reference:
  • Yang Zhang, I-TASSER: Fully automated protein structure prediction in CASP8. Proteins: Structure, Function, and Bioinformatics, 77 (Suppl 9): 100-113 (2009). (Download the PDF file and Support Information).
• The CASP9 Decoy Set is a decoy set generated by I-TASSER and QUARK for 85 single-domain targets in the CASP9 experiment. Each target contains around 200-780 decoys, from which all the Zhang-Server and QUARK server models have been selected.
  Reference:
  • Dong Xu, Jian Zhang, Ambrish Roy, Yang Zhang. Automated protein structure modeling in CASP9 by I-TASSER pipeline combined with QUARK-based ab initio folding and FG-MD-based structure refinement. Proteins: Structure, Function, and Bioinformatics, 2011, 79 (Suppl 10) 147-160 (download the PDF file).
• The CASP10 Decoy Set is a decoy set generated by I-TASSER and QUARK for 43 single-domain targets in the CASP10 experiment. Each target contains around 200-825 decoys, from which all the Zhang-Server and QUARK server models have been selected.
  Reference:
  • Yang Zhang, Interplay of I-TASSER and QUARK for template-based and ab initio protein structure prediction in CASP10. Proteins: Structure, Function, and Bioinformatics, 2014, 82 (Suppl 2), 175-187 (download the PDF file and Support Information).
• The CASP11 Decoy Set is a decoy set generated by I-TASSER and QUARK for 64 single-domain targets in the CASP11 experiment. Each target contains around 400-1500 decoys, from which all the Zhang-Server and QUARK server models have been selected.