Protein-protein binding affinity prediction based on an SVR ensemble

Xueling Li, Min Zhu, Xiaolai Li, Hong Qiang Wang, Shulin Wang

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

9 Citations (Scopus)

Abstract

Accurately predicting generic protein-protein binding affinities (PPBA) is essential to analyze the outputs of protein docking and may help infer real status of cellular protein-protein interaction sub-networks. However, accurate PPBA prediction is still extremely challenging. Machine learning methods are promising to address this problem. We propose a two-layer support vector regression (TLSVR) model to implicitly capture binding contributions that are hard to explicitly model. The TLSVR circumvents both the descriptor compatibility problem and the need for problematic modeling assumptions. Input features for TLSVR in first layer are scores of 2209 interacting atom pairs within each distance bin. The base SVRs are combined by the second layer to infer the final affinities. Leave-one-out validation on our heterogeneous data shows that the TLSVR method obtains a very good result of R=0.80 and SD=1.32 with real affinities. Comparison experiment further demonstrates that TLSVR is superior to the previous state-of-art methods in predicting generic PPBA.

Original languageEnglish (US)
Title of host publicationIntelligent Computing Technology - 8th International Conference, ICIC 2012, Proceedings
Pages145-151
Number of pages7
Volume7389 LNCS
DOIs
StatePublished - 2012
Externally publishedYes
Event8th International Conference on Intelligent Computing Technology, ICIC 2012 - Huangshan, China
Duration: Jul 25 2012Jul 29 2012

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume7389 LNCS
ISSN (Print)03029743
ISSN (Electronic)16113349

Other

Other8th International Conference on Intelligent Computing Technology, ICIC 2012
CountryChina
CityHuangshan
Period7/25/127/29/12

Fingerprint

Affine transformation
Support Vector Regression
Ensemble
Proteins
Protein
Prediction
Bins
Docking
Protein-protein Interaction
Learning systems
Protein Binding
Compatibility
Descriptors
Regression Model
Machine Learning
Atoms
Output
Modeling
Demonstrate
Experiment

Keywords

  • machine learning
  • potential of mean force
  • Protein-protein interaction affinity
  • two-layer support vector machine

ASJC Scopus subject areas

  • Computer Science(all)
  • Theoretical Computer Science

Cite this

Li, X., Zhu, M., Li, X., Wang, H. Q., & Wang, S. (2012). Protein-protein binding affinity prediction based on an SVR ensemble. In Intelligent Computing Technology - 8th International Conference, ICIC 2012, Proceedings (Vol. 7389 LNCS, pp. 145-151). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7389 LNCS). https://doi.org/10.1007/978-3-642-31588-6_19

Protein-protein binding affinity prediction based on an SVR ensemble. / Li, Xueling; Zhu, Min; Li, Xiaolai; Wang, Hong Qiang; Wang, Shulin.

Intelligent Computing Technology - 8th International Conference, ICIC 2012, Proceedings. Vol. 7389 LNCS 2012. p. 145-151 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 7389 LNCS).

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

Li, X, Zhu, M, Li, X, Wang, HQ & Wang, S 2012, Protein-protein binding affinity prediction based on an SVR ensemble. in Intelligent Computing Technology - 8th International Conference, ICIC 2012, Proceedings. vol. 7389 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 7389 LNCS, pp. 145-151, 8th International Conference on Intelligent Computing Technology, ICIC 2012, Huangshan, China, 7/25/12. https://doi.org/10.1007/978-3-642-31588-6_19
Li X, Zhu M, Li X, Wang HQ, Wang S. Protein-protein binding affinity prediction based on an SVR ensemble. In Intelligent Computing Technology - 8th International Conference, ICIC 2012, Proceedings. Vol. 7389 LNCS. 2012. p. 145-151. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-642-31588-6_19
Li, Xueling ; Zhu, Min ; Li, Xiaolai ; Wang, Hong Qiang ; Wang, Shulin. / Protein-protein binding affinity prediction based on an SVR ensemble. Intelligent Computing Technology - 8th International Conference, ICIC 2012, Proceedings. Vol. 7389 LNCS 2012. pp. 145-151 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{75e00ed3bd994ebda8b0b62621a7d1d1,
title = "Protein-protein binding affinity prediction based on an SVR ensemble",
abstract = "Accurately predicting generic protein-protein binding affinities (PPBA) is essential to analyze the outputs of protein docking and may help infer real status of cellular protein-protein interaction sub-networks. However, accurate PPBA prediction is still extremely challenging. Machine learning methods are promising to address this problem. We propose a two-layer support vector regression (TLSVR) model to implicitly capture binding contributions that are hard to explicitly model. The TLSVR circumvents both the descriptor compatibility problem and the need for problematic modeling assumptions. Input features for TLSVR in first layer are scores of 2209 interacting atom pairs within each distance bin. The base SVRs are combined by the second layer to infer the final affinities. Leave-one-out validation on our heterogeneous data shows that the TLSVR method obtains a very good result of R=0.80 and SD=1.32 with real affinities. Comparison experiment further demonstrates that TLSVR is superior to the previous state-of-art methods in predicting generic PPBA.",
keywords = "machine learning, potential of mean force, Protein-protein interaction affinity, two-layer support vector machine",
author = "Xueling Li and Min Zhu and Xiaolai Li and Wang, {Hong Qiang} and Shulin Wang",
year = "2012",
doi = "10.1007/978-3-642-31588-6_19",
language = "English (US)",
isbn = "9783642315879",
volume = "7389 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
pages = "145--151",
booktitle = "Intelligent Computing Technology - 8th International Conference, ICIC 2012, Proceedings",

}

TY - GEN

T1 - Protein-protein binding affinity prediction based on an SVR ensemble

AU - Li, Xueling

AU - Zhu, Min

AU - Li, Xiaolai

AU - Wang, Hong Qiang

AU - Wang, Shulin

PY - 2012

Y1 - 2012

N2 - Accurately predicting generic protein-protein binding affinities (PPBA) is essential to analyze the outputs of protein docking and may help infer real status of cellular protein-protein interaction sub-networks. However, accurate PPBA prediction is still extremely challenging. Machine learning methods are promising to address this problem. We propose a two-layer support vector regression (TLSVR) model to implicitly capture binding contributions that are hard to explicitly model. The TLSVR circumvents both the descriptor compatibility problem and the need for problematic modeling assumptions. Input features for TLSVR in first layer are scores of 2209 interacting atom pairs within each distance bin. The base SVRs are combined by the second layer to infer the final affinities. Leave-one-out validation on our heterogeneous data shows that the TLSVR method obtains a very good result of R=0.80 and SD=1.32 with real affinities. Comparison experiment further demonstrates that TLSVR is superior to the previous state-of-art methods in predicting generic PPBA.

AB - Accurately predicting generic protein-protein binding affinities (PPBA) is essential to analyze the outputs of protein docking and may help infer real status of cellular protein-protein interaction sub-networks. However, accurate PPBA prediction is still extremely challenging. Machine learning methods are promising to address this problem. We propose a two-layer support vector regression (TLSVR) model to implicitly capture binding contributions that are hard to explicitly model. The TLSVR circumvents both the descriptor compatibility problem and the need for problematic modeling assumptions. Input features for TLSVR in first layer are scores of 2209 interacting atom pairs within each distance bin. The base SVRs are combined by the second layer to infer the final affinities. Leave-one-out validation on our heterogeneous data shows that the TLSVR method obtains a very good result of R=0.80 and SD=1.32 with real affinities. Comparison experiment further demonstrates that TLSVR is superior to the previous state-of-art methods in predicting generic PPBA.

KW - machine learning

KW - potential of mean force

KW - Protein-protein interaction affinity

KW - two-layer support vector machine

UR - http://www.scopus.com/inward/record.url?scp=84865305788&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84865305788&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-31588-6_19

DO - 10.1007/978-3-642-31588-6_19

M3 - Conference contribution

AN - SCOPUS:84865305788

SN - 9783642315879

VL - 7389 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 145

EP - 151

BT - Intelligent Computing Technology - 8th International Conference, ICIC 2012, Proceedings

ER -