Compressed images for affinity prediction (CIFAP): a study on predicting binding affinities for checkpoint kinase 1 protein inhibitors

Abstract

Analyses of known protein–ligand interactions play an important role in designing novel and efficient drugs, contributing to drug discovery and development. Recently, machine learning methods have proven useful in the design of novel drugs, which utilize intelligent techniques to predict the outcome of unknown protein–ligand interactions by learning from the physical and geometrical properties of known protein–ligand interactions. The aim of this study is to work through a specific example of a novel computational method, namely compressed images for affinity prediction (CIFAP), in which binding affinities for structurally related ligands in complexes with human checkpoint kinase 1 (CHK1) are predicted. The CIFAP algorithm presented here relates published pIC 50 values of 57 compounds, derived from a thienopyridine pharmacophore, in complexes with CHK1 to their two‐dimensional (2D) electrostatic potential images compressed in orthogonal dimensions. Patterns obtained from the 2D images are then used as inputs in regression and learning algorithms such as support vector regression (SVR) and adaptive neuro‐fuzzy inference system (ANFIS) methods to validate the experimental pIC 50 values. This study revealed that the 2D image pixels in the vicinity of bound ligand surfaces provide more relevant information to make correlations with the empirical pIC 50 values. As compared with ANFIS, SVR gave rise to the lowest root mean square errors and the greatest correlations, suggesting that SVR could be a plausible choice of machine learning methods in predicting binding affinities by CIFAP. Copyright © 2013 John Wiley & Sons, Ltd.

Compressed images for affinity prediction-2 (CIFAP-2): an improved machine learning methodology on protein–ligand interactions based on a study on caspase 3 inhibitors

“…In this study, three prediction methods were applied by MATLAB 31 to test CIFAP-2 on CASP3 inhibitors: (1) PLSR 17 , (2) Support Vector Regression (SVR) 14 , and (3) Adaptive NeuroFuzzy Inference System (ANFIS) 15 . Theoretical background for the first prediction method is given in details in the following subsection, while the last two prediction methods were already described by Erdaş et al 12 as prediction methods utilized in the development of CIFAP.…”

“…In general, future sets are grouped into training and test sets for validation by, for instance, repeat random subsampling (RRS) 12,33 . However, RRS is not applied in this study for validation of CASP3-ligand complexes due to insufficient number of the ligands.…”

“…General performance terms used in this study for the evaluation of predicted pIC 50 values are root mean square error (RMSE) and coeffcient of determination (R 2 ), whose equations and their implementation were previously described by Erdas et al 12 .…”

“…It is possible to search for or predict specific properties of new drugs by using information from known ligand-receptor interactions in association with bioinformatics and machine learning methods 11 . We previously reported 12 a novel pharmacophore-based drug development algorithm for data representation, namely Compressed Images for Binding Affnity Prediction (CIFAP), to predict binding affinities for an array of different ligands (carrying a common pharmacophore) interacting with the binding site of a protein. CIFAP is implemented in two phases: In the first phase, CIFAP involves preperation of orthogonal 2D-compressed grid images of experimentally determined or docked coordinates of ligand-receptor complexes.…”

“…In this study, CIFAP-2, an improved version of CIFAP 12 , is proposed based on some fine-tunings applied on two main phases taken to implement CIFAP. While the first phase of CIFAP-2 is identical to that of CIFAP, the second phase of CIFAP-2 utilizes the sequential forward floating selection (SFFS) method 16 , but not SFS as in CIFAP.…”