Prediction of peptide bonding affinity: kernel methods for nonlinear modeling

TL;DR

This paper evaluates kernel methods for predicting peptide binding affinity, demonstrating that nonlinear models like kernel PLS outperform linear approaches and that transferable atom features enhance prediction accuracy.

Contribution

It introduces the application of kernel partial least squares and transferable atom features for improved peptide binding affinity prediction.

Findings

Kernel PLS outperforms traditional PLS.

Transferable atom features improve predictive accuracy.

Nonlinear modeling enhances peptide affinity prediction.

Abstract

This paper presents regression models obtained from a process of blind prediction of peptide binding affinity from provided descriptors for several distinct datasets as part of the 2006 Comparative Evaluation of Prediction Algorithms (COEPRA) contest. This paper finds that kernel partial least squares, a nonlinear partial least squares (PLS) algorithm, outperforms PLS, and that the incorporation of transferable atom equivalent features improves predictive capability.