GemAffinity

A scoring function for predicting binding affinity and virtual screening

Kai Cheng Hsu, Yen Fu Chen, Jinn Moon Yang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Prediction of protein-ligand binding affinities is an important issue in molecular recognition and virtual screening. We have developed a scoring function, namely GemAffinity, to predict binding affinities by analyzing 88 descriptors derived from 891 protein-ligand structures selected from the Protein Data Bank (PDB). Based on these 88 descriptors, we derived GemAffinity using a stepwise regression method to identify five descriptors, including van der Waals contact; metal-ligand interactions; water effects; ligand deformation penalties; and highly conserved residues interacting to a bound ligand with hydrogen bonds. GemAffinity was evaluated on an independent set, and the correlation between predicted and experimental values is 0.572. GemAffinity is the best among 13 methods on this set. Our GemAffinity was then applied to virtual screening for thymidine kinase (TK), human carbonic anhydrase II (HCAII), estrogen receptor of antagonists (ER) and agonists (ERA). Experimental results indicate that GemAffinity is able to reduce the disadvantages (i.e. preferring highly polar or high molecular weight compounds) of energy-based scoring functions. In addition, GemAffinity easily combined with other scoring functions to enrich screening accuracies. We believe that GemAffinity is useful to predict binding affinity and virtual screening.

Original languageEnglish
Title of host publication2009 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2009
Pages309-314
Number of pages6
DOIs
Publication statusPublished - 2009
Externally publishedYes
Event2009 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2009 - Washington, D.C., United States
Duration: Nov 1 2009Nov 4 2009

Other

Other2009 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2009
CountryUnited States
CityWashington, D.C.
Period11/1/0911/4/09

Fingerprint

Screening
Ligands
Proteins
Carbonic anhydrase
Carbonic Anhydrase II
Molecular recognition
Thymidine Kinase
Protein Binding
Hydrogen
Hydrogen bonds
Estrogens
Molecular Weight
Metals
Molecular weight
Databases
Water

Keywords

  • Binding affinity prediction
  • Component
  • Protein-ligand interactions
  • Scoring functions
  • Structure-based drug design
  • Virtual screening

ASJC Scopus subject areas

  • Artificial Intelligence
  • Software
  • Biomedical Engineering
  • Health Informatics

Cite this

Hsu, K. C., Chen, Y. F., & Yang, J. M. (2009). GemAffinity: A scoring function for predicting binding affinity and virtual screening. In 2009 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2009 (pp. 309-314). [5341774] https://doi.org/10.1109/BIBM.2009.24

GemAffinity : A scoring function for predicting binding affinity and virtual screening. / Hsu, Kai Cheng; Chen, Yen Fu; Yang, Jinn Moon.

2009 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2009. 2009. p. 309-314 5341774.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hsu, KC, Chen, YF & Yang, JM 2009, GemAffinity: A scoring function for predicting binding affinity and virtual screening. in 2009 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2009., 5341774, pp. 309-314, 2009 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2009, Washington, D.C., United States, 11/1/09. https://doi.org/10.1109/BIBM.2009.24
Hsu KC, Chen YF, Yang JM. GemAffinity: A scoring function for predicting binding affinity and virtual screening. In 2009 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2009. 2009. p. 309-314. 5341774 https://doi.org/10.1109/BIBM.2009.24
Hsu, Kai Cheng ; Chen, Yen Fu ; Yang, Jinn Moon. / GemAffinity : A scoring function for predicting binding affinity and virtual screening. 2009 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2009. 2009. pp. 309-314
@inproceedings{c795476cb3dc4dfda9e01e2d9f8537ac,
title = "GemAffinity: A scoring function for predicting binding affinity and virtual screening",
abstract = "Prediction of protein-ligand binding affinities is an important issue in molecular recognition and virtual screening. We have developed a scoring function, namely GemAffinity, to predict binding affinities by analyzing 88 descriptors derived from 891 protein-ligand structures selected from the Protein Data Bank (PDB). Based on these 88 descriptors, we derived GemAffinity using a stepwise regression method to identify five descriptors, including van der Waals contact; metal-ligand interactions; water effects; ligand deformation penalties; and highly conserved residues interacting to a bound ligand with hydrogen bonds. GemAffinity was evaluated on an independent set, and the correlation between predicted and experimental values is 0.572. GemAffinity is the best among 13 methods on this set. Our GemAffinity was then applied to virtual screening for thymidine kinase (TK), human carbonic anhydrase II (HCAII), estrogen receptor of antagonists (ER) and agonists (ERA). Experimental results indicate that GemAffinity is able to reduce the disadvantages (i.e. preferring highly polar or high molecular weight compounds) of energy-based scoring functions. In addition, GemAffinity easily combined with other scoring functions to enrich screening accuracies. We believe that GemAffinity is useful to predict binding affinity and virtual screening.",
keywords = "Binding affinity prediction, Component, Protein-ligand interactions, Scoring functions, Structure-based drug design, Virtual screening",
author = "Hsu, {Kai Cheng} and Chen, {Yen Fu} and Yang, {Jinn Moon}",
year = "2009",
doi = "10.1109/BIBM.2009.24",
language = "English",
isbn = "9780769538853",
pages = "309--314",
booktitle = "2009 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2009",

}

TY - GEN

T1 - GemAffinity

T2 - A scoring function for predicting binding affinity and virtual screening

AU - Hsu, Kai Cheng

AU - Chen, Yen Fu

AU - Yang, Jinn Moon

PY - 2009

Y1 - 2009

N2 - Prediction of protein-ligand binding affinities is an important issue in molecular recognition and virtual screening. We have developed a scoring function, namely GemAffinity, to predict binding affinities by analyzing 88 descriptors derived from 891 protein-ligand structures selected from the Protein Data Bank (PDB). Based on these 88 descriptors, we derived GemAffinity using a stepwise regression method to identify five descriptors, including van der Waals contact; metal-ligand interactions; water effects; ligand deformation penalties; and highly conserved residues interacting to a bound ligand with hydrogen bonds. GemAffinity was evaluated on an independent set, and the correlation between predicted and experimental values is 0.572. GemAffinity is the best among 13 methods on this set. Our GemAffinity was then applied to virtual screening for thymidine kinase (TK), human carbonic anhydrase II (HCAII), estrogen receptor of antagonists (ER) and agonists (ERA). Experimental results indicate that GemAffinity is able to reduce the disadvantages (i.e. preferring highly polar or high molecular weight compounds) of energy-based scoring functions. In addition, GemAffinity easily combined with other scoring functions to enrich screening accuracies. We believe that GemAffinity is useful to predict binding affinity and virtual screening.

AB - Prediction of protein-ligand binding affinities is an important issue in molecular recognition and virtual screening. We have developed a scoring function, namely GemAffinity, to predict binding affinities by analyzing 88 descriptors derived from 891 protein-ligand structures selected from the Protein Data Bank (PDB). Based on these 88 descriptors, we derived GemAffinity using a stepwise regression method to identify five descriptors, including van der Waals contact; metal-ligand interactions; water effects; ligand deformation penalties; and highly conserved residues interacting to a bound ligand with hydrogen bonds. GemAffinity was evaluated on an independent set, and the correlation between predicted and experimental values is 0.572. GemAffinity is the best among 13 methods on this set. Our GemAffinity was then applied to virtual screening for thymidine kinase (TK), human carbonic anhydrase II (HCAII), estrogen receptor of antagonists (ER) and agonists (ERA). Experimental results indicate that GemAffinity is able to reduce the disadvantages (i.e. preferring highly polar or high molecular weight compounds) of energy-based scoring functions. In addition, GemAffinity easily combined with other scoring functions to enrich screening accuracies. We believe that GemAffinity is useful to predict binding affinity and virtual screening.

KW - Binding affinity prediction

KW - Component

KW - Protein-ligand interactions

KW - Scoring functions

KW - Structure-based drug design

KW - Virtual screening

UR - http://www.scopus.com/inward/record.url?scp=74549159654&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=74549159654&partnerID=8YFLogxK

U2 - 10.1109/BIBM.2009.24

DO - 10.1109/BIBM.2009.24

M3 - Conference contribution

SN - 9780769538853

SP - 309

EP - 314

BT - 2009 IEEE International Conference on Bioinformatics and Biomedicine, BIBM 2009

ER -