Predicting binding affinity of CSAR ligands using both structure-based and ligand-based approaches

Denis Fourches, Eugene Muratov, Feng Ding, Nikolay Dokholyan, Alexander Tropsha

Research output: Contribution to journalArticle

17 Citations (Scopus)

Abstract

We report on the prediction accuracy of ligand-based (2D QSAR) and structure-based (MedusaDock) methods used both independently and in consensus for ranking the congeneric series of ligands binding to three protein targets (UK, ERK2, and CHK1) from the CSAR 2011 benchmark exercise. An ensemble of predictive QSAR models was developed using known binders of these three targets extracted from the publicly available ChEMBL database. Selected models were used to predict the binding affinity of CSAR compounds toward the corresponding targets and rank them accordingly; the overall ranking accuracy evaluated by Spearman correlation was as high as 0.78 for UK, 0.60 for ERK2, and 0.56 for CHK1, placing our predictions in the top 10% among all the participants. In parallel, MedusaDock, designed to predict reliable docking poses, was also used for ranking the CSAR ligands according to their docking scores; the resulting accuracy (Spearman correlation) for UK, ERK2, and CHK1 were 0.76, 0.31, and 0.26, respectively. In addition, performance of several consensus approaches combining MedusaDock- and QSAR-predicted ranks altogether has been explored; the best approach yielded Spearman correlation coefficients for UK, ERK2, and CHK1 of 0.82, 0.50, and 0.45, respectively. This study shows that (i) externally validated 2D QSAR models were capable of ranking CSAR ligands at least as accurately as more computationally intensive structure-based approaches used both by us and by other groups and (ii) ligand-based QSAR models can complement structure-based approaches by boosting the prediction performances when used in consensus.

Original languageEnglish (US)
Pages (from-to)1915-1922
Number of pages8
JournalJournal of Chemical Information and Modeling
Volume53
Issue number8
DOIs
StatePublished - Aug 26 2013

Fingerprint

ranking
Ligands
predictive model
performance
Binders
Proteins
Group

All Science Journal Classification (ASJC) codes

  • Chemistry(all)
  • Chemical Engineering(all)
  • Computer Science Applications
  • Library and Information Sciences

Cite this

Fourches, Denis ; Muratov, Eugene ; Ding, Feng ; Dokholyan, Nikolay ; Tropsha, Alexander. / Predicting binding affinity of CSAR ligands using both structure-based and ligand-based approaches. In: Journal of Chemical Information and Modeling. 2013 ; Vol. 53, No. 8. pp. 1915-1922.
@article{de3eacd8a3af49838fc66a1db16d6771,
title = "Predicting binding affinity of CSAR ligands using both structure-based and ligand-based approaches",
abstract = "We report on the prediction accuracy of ligand-based (2D QSAR) and structure-based (MedusaDock) methods used both independently and in consensus for ranking the congeneric series of ligands binding to three protein targets (UK, ERK2, and CHK1) from the CSAR 2011 benchmark exercise. An ensemble of predictive QSAR models was developed using known binders of these three targets extracted from the publicly available ChEMBL database. Selected models were used to predict the binding affinity of CSAR compounds toward the corresponding targets and rank them accordingly; the overall ranking accuracy evaluated by Spearman correlation was as high as 0.78 for UK, 0.60 for ERK2, and 0.56 for CHK1, placing our predictions in the top 10{\%} among all the participants. In parallel, MedusaDock, designed to predict reliable docking poses, was also used for ranking the CSAR ligands according to their docking scores; the resulting accuracy (Spearman correlation) for UK, ERK2, and CHK1 were 0.76, 0.31, and 0.26, respectively. In addition, performance of several consensus approaches combining MedusaDock- and QSAR-predicted ranks altogether has been explored; the best approach yielded Spearman correlation coefficients for UK, ERK2, and CHK1 of 0.82, 0.50, and 0.45, respectively. This study shows that (i) externally validated 2D QSAR models were capable of ranking CSAR ligands at least as accurately as more computationally intensive structure-based approaches used both by us and by other groups and (ii) ligand-based QSAR models can complement structure-based approaches by boosting the prediction performances when used in consensus.",
author = "Denis Fourches and Eugene Muratov and Feng Ding and Nikolay Dokholyan and Alexander Tropsha",
year = "2013",
month = "8",
day = "26",
doi = "10.1021/ci400216q",
language = "English (US)",
volume = "53",
pages = "1915--1922",
journal = "Journal of Chemical Information and Modeling",
issn = "1549-9596",
publisher = "American Chemical Society",
number = "8",

}

Predicting binding affinity of CSAR ligands using both structure-based and ligand-based approaches. / Fourches, Denis; Muratov, Eugene; Ding, Feng; Dokholyan, Nikolay; Tropsha, Alexander.

In: Journal of Chemical Information and Modeling, Vol. 53, No. 8, 26.08.2013, p. 1915-1922.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Predicting binding affinity of CSAR ligands using both structure-based and ligand-based approaches

AU - Fourches, Denis

AU - Muratov, Eugene

AU - Ding, Feng

AU - Dokholyan, Nikolay

AU - Tropsha, Alexander

PY - 2013/8/26

Y1 - 2013/8/26

N2 - We report on the prediction accuracy of ligand-based (2D QSAR) and structure-based (MedusaDock) methods used both independently and in consensus for ranking the congeneric series of ligands binding to three protein targets (UK, ERK2, and CHK1) from the CSAR 2011 benchmark exercise. An ensemble of predictive QSAR models was developed using known binders of these three targets extracted from the publicly available ChEMBL database. Selected models were used to predict the binding affinity of CSAR compounds toward the corresponding targets and rank them accordingly; the overall ranking accuracy evaluated by Spearman correlation was as high as 0.78 for UK, 0.60 for ERK2, and 0.56 for CHK1, placing our predictions in the top 10% among all the participants. In parallel, MedusaDock, designed to predict reliable docking poses, was also used for ranking the CSAR ligands according to their docking scores; the resulting accuracy (Spearman correlation) for UK, ERK2, and CHK1 were 0.76, 0.31, and 0.26, respectively. In addition, performance of several consensus approaches combining MedusaDock- and QSAR-predicted ranks altogether has been explored; the best approach yielded Spearman correlation coefficients for UK, ERK2, and CHK1 of 0.82, 0.50, and 0.45, respectively. This study shows that (i) externally validated 2D QSAR models were capable of ranking CSAR ligands at least as accurately as more computationally intensive structure-based approaches used both by us and by other groups and (ii) ligand-based QSAR models can complement structure-based approaches by boosting the prediction performances when used in consensus.

AB - We report on the prediction accuracy of ligand-based (2D QSAR) and structure-based (MedusaDock) methods used both independently and in consensus for ranking the congeneric series of ligands binding to three protein targets (UK, ERK2, and CHK1) from the CSAR 2011 benchmark exercise. An ensemble of predictive QSAR models was developed using known binders of these three targets extracted from the publicly available ChEMBL database. Selected models were used to predict the binding affinity of CSAR compounds toward the corresponding targets and rank them accordingly; the overall ranking accuracy evaluated by Spearman correlation was as high as 0.78 for UK, 0.60 for ERK2, and 0.56 for CHK1, placing our predictions in the top 10% among all the participants. In parallel, MedusaDock, designed to predict reliable docking poses, was also used for ranking the CSAR ligands according to their docking scores; the resulting accuracy (Spearman correlation) for UK, ERK2, and CHK1 were 0.76, 0.31, and 0.26, respectively. In addition, performance of several consensus approaches combining MedusaDock- and QSAR-predicted ranks altogether has been explored; the best approach yielded Spearman correlation coefficients for UK, ERK2, and CHK1 of 0.82, 0.50, and 0.45, respectively. This study shows that (i) externally validated 2D QSAR models were capable of ranking CSAR ligands at least as accurately as more computationally intensive structure-based approaches used both by us and by other groups and (ii) ligand-based QSAR models can complement structure-based approaches by boosting the prediction performances when used in consensus.

UR - http://www.scopus.com/inward/record.url?scp=84883217252&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84883217252&partnerID=8YFLogxK

U2 - 10.1021/ci400216q

DO - 10.1021/ci400216q

M3 - Article

C2 - 23809015

AN - SCOPUS:84883217252

VL - 53

SP - 1915

EP - 1922

JO - Journal of Chemical Information and Modeling

JF - Journal of Chemical Information and Modeling

SN - 1549-9596

IS - 8

ER -