An accurate, residue-level, pair potential of mean force for folding and binding based on the distance-scaled, ideal-gas reference state

Chi Zhang, Song Liu, Hongyi Zhou, Yaoqi Zhou

Research output: Contribution to journalArticle

135 Citations (Scopus)

Abstract

Structure prediction on a genomic scale requires a simplified energy function that can efficiently sample the conformational space of polypeptide chains. A good energy function at minimum should discriminate native structures against decoys. Here, we show that a recently developed, residue-specific, all-atom knowledge-based potential (167 atomic types) based on distance-scaled, finite ideal-gas reference state (DFIRE-all-atom) can be substantially simplified to 20 residue types located at side-chain center of mass (DFIRE-SCM) without a significant change in its capability of structure discrimination. Using 96 standard multiple decoy sets, we show that there is only a small reduction (from 80% to 78%) in success rate of ranking native structures as the top 1. The success rate is higher than two previously developed, all-atom distance-dependent statistical pair potentials. Applied to structure selections of 21 docking decoys without modification, the DFIRE-SCM potential is 29% more successful in recognizing native complex structures than an all-atom statistical potential trained by a database of dimeric interfaces. The potential also achieves 92% accuracy in distinguishing true dimeric interfaces from artificial crystal interfaces. In addition, the DFIRE potential with the Cα positions as the interaction centers recognizes 123 native structures out of a comprehensive 125-protein TOUCHSTONE decoy set in which each protein has 24,000 decoys with only Cα positions. Furthermore, the performance by DFIRE-SCM on newly established 25 monomeric and 31 docking Rosetta-decoy sets is comparable to (or better than in the case of monomeric decoy sets) that of a recently developed, all-atom Rosetta energy function enhanced with an orientation-dependent hydrogen bonding potential.

Original languageEnglish (US)
Pages (from-to)400-411
Number of pages12
JournalProtein Science
Volume13
Issue number2
DOIs
StatePublished - Feb 1 2004

Fingerprint

Gases
Atoms
Hydrogen Bonding
Proteins
Databases
Peptides
Hydrogen bonds
Crystals

Keywords

  • Decoy sets
  • Ideal-gas reference state
  • Knowledge-based potential
  • Residue-level potential

ASJC Scopus subject areas

  • Biochemistry
  • Molecular Biology

Cite this

An accurate, residue-level, pair potential of mean force for folding and binding based on the distance-scaled, ideal-gas reference state. / Zhang, Chi; Liu, Song; Zhou, Hongyi; Zhou, Yaoqi.

In: Protein Science, Vol. 13, No. 2, 01.02.2004, p. 400-411.

Research output: Contribution to journalArticle

@article{df5218656126409e86f54b9fd6104c5b,
title = "An accurate, residue-level, pair potential of mean force for folding and binding based on the distance-scaled, ideal-gas reference state",
abstract = "Structure prediction on a genomic scale requires a simplified energy function that can efficiently sample the conformational space of polypeptide chains. A good energy function at minimum should discriminate native structures against decoys. Here, we show that a recently developed, residue-specific, all-atom knowledge-based potential (167 atomic types) based on distance-scaled, finite ideal-gas reference state (DFIRE-all-atom) can be substantially simplified to 20 residue types located at side-chain center of mass (DFIRE-SCM) without a significant change in its capability of structure discrimination. Using 96 standard multiple decoy sets, we show that there is only a small reduction (from 80{\%} to 78{\%}) in success rate of ranking native structures as the top 1. The success rate is higher than two previously developed, all-atom distance-dependent statistical pair potentials. Applied to structure selections of 21 docking decoys without modification, the DFIRE-SCM potential is 29{\%} more successful in recognizing native complex structures than an all-atom statistical potential trained by a database of dimeric interfaces. The potential also achieves 92{\%} accuracy in distinguishing true dimeric interfaces from artificial crystal interfaces. In addition, the DFIRE potential with the Cα positions as the interaction centers recognizes 123 native structures out of a comprehensive 125-protein TOUCHSTONE decoy set in which each protein has 24,000 decoys with only Cα positions. Furthermore, the performance by DFIRE-SCM on newly established 25 monomeric and 31 docking Rosetta-decoy sets is comparable to (or better than in the case of monomeric decoy sets) that of a recently developed, all-atom Rosetta energy function enhanced with an orientation-dependent hydrogen bonding potential.",
keywords = "Decoy sets, Ideal-gas reference state, Knowledge-based potential, Residue-level potential",
author = "Chi Zhang and Song Liu and Hongyi Zhou and Yaoqi Zhou",
year = "2004",
month = "2",
day = "1",
doi = "10.1110/ps.03348304",
language = "English (US)",
volume = "13",
pages = "400--411",
journal = "Protein Science",
issn = "0961-8368",
publisher = "Cold Spring Harbor Laboratory Press",
number = "2",

}

TY - JOUR

T1 - An accurate, residue-level, pair potential of mean force for folding and binding based on the distance-scaled, ideal-gas reference state

AU - Zhang, Chi

AU - Liu, Song

AU - Zhou, Hongyi

AU - Zhou, Yaoqi

PY - 2004/2/1

Y1 - 2004/2/1

N2 - Structure prediction on a genomic scale requires a simplified energy function that can efficiently sample the conformational space of polypeptide chains. A good energy function at minimum should discriminate native structures against decoys. Here, we show that a recently developed, residue-specific, all-atom knowledge-based potential (167 atomic types) based on distance-scaled, finite ideal-gas reference state (DFIRE-all-atom) can be substantially simplified to 20 residue types located at side-chain center of mass (DFIRE-SCM) without a significant change in its capability of structure discrimination. Using 96 standard multiple decoy sets, we show that there is only a small reduction (from 80% to 78%) in success rate of ranking native structures as the top 1. The success rate is higher than two previously developed, all-atom distance-dependent statistical pair potentials. Applied to structure selections of 21 docking decoys without modification, the DFIRE-SCM potential is 29% more successful in recognizing native complex structures than an all-atom statistical potential trained by a database of dimeric interfaces. The potential also achieves 92% accuracy in distinguishing true dimeric interfaces from artificial crystal interfaces. In addition, the DFIRE potential with the Cα positions as the interaction centers recognizes 123 native structures out of a comprehensive 125-protein TOUCHSTONE decoy set in which each protein has 24,000 decoys with only Cα positions. Furthermore, the performance by DFIRE-SCM on newly established 25 monomeric and 31 docking Rosetta-decoy sets is comparable to (or better than in the case of monomeric decoy sets) that of a recently developed, all-atom Rosetta energy function enhanced with an orientation-dependent hydrogen bonding potential.

AB - Structure prediction on a genomic scale requires a simplified energy function that can efficiently sample the conformational space of polypeptide chains. A good energy function at minimum should discriminate native structures against decoys. Here, we show that a recently developed, residue-specific, all-atom knowledge-based potential (167 atomic types) based on distance-scaled, finite ideal-gas reference state (DFIRE-all-atom) can be substantially simplified to 20 residue types located at side-chain center of mass (DFIRE-SCM) without a significant change in its capability of structure discrimination. Using 96 standard multiple decoy sets, we show that there is only a small reduction (from 80% to 78%) in success rate of ranking native structures as the top 1. The success rate is higher than two previously developed, all-atom distance-dependent statistical pair potentials. Applied to structure selections of 21 docking decoys without modification, the DFIRE-SCM potential is 29% more successful in recognizing native complex structures than an all-atom statistical potential trained by a database of dimeric interfaces. The potential also achieves 92% accuracy in distinguishing true dimeric interfaces from artificial crystal interfaces. In addition, the DFIRE potential with the Cα positions as the interaction centers recognizes 123 native structures out of a comprehensive 125-protein TOUCHSTONE decoy set in which each protein has 24,000 decoys with only Cα positions. Furthermore, the performance by DFIRE-SCM on newly established 25 monomeric and 31 docking Rosetta-decoy sets is comparable to (or better than in the case of monomeric decoy sets) that of a recently developed, all-atom Rosetta energy function enhanced with an orientation-dependent hydrogen bonding potential.

KW - Decoy sets

KW - Ideal-gas reference state

KW - Knowledge-based potential

KW - Residue-level potential

UR - http://www.scopus.com/inward/record.url?scp=1642534609&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=1642534609&partnerID=8YFLogxK

U2 - 10.1110/ps.03348304

DO - 10.1110/ps.03348304

M3 - Article

C2 - 14739325

AN - SCOPUS:1642534609

VL - 13

SP - 400

EP - 411

JO - Protein Science

JF - Protein Science

SN - 0961-8368

IS - 2

ER -