CSCORE: A SIMPLE YET EFFECTIVE SCORING FUNCTION FOR PROTEIN–LIGAND BINDING AFFINITY PREDICTION USING MODIFIED CMAC LEARNING ARCHITECTURE

XUCHANG OUYANG, STEPHANUS DANIEL HANDOKO, CHEE KEONG KWOH
2011 Journal of Bioinformatics and Computational Biology  
Protein-ligand docking is a computational method to identify the binding mode of a 18 ligand and a target protein, and predict the corresponding binding affinity using a scor-19 ing function. This method has great value in drug design. After decades of development, 20 scoring functions nowadays typically can identify the true binding mode, but the pre-21 diction of binding affinity still remains a major problem. Here we present CScore, a 22 data-driven scoring function using a modified
more » ... a modified Cerebellar Model Articulation Controller 23 (CMAC) learning architecture, for accurate binding affinity prediction. The performance 24 of CScore in terms of correlation between predicted and experimental binding affinities 25 is benchmarked under different validation approaches. CScore achieves a prediction with 26 R = 0.7668 and RMSE = 1.4540 when tested on an independent data set. To the best of 27 our knowledge, this result outperforms other scoring functions tested on the same data 28 set. The performance of CScore varies on different clusters under the leave-cluster-out 29 validation approach, but still achieves competitive result. Lastly, the target-specified 30 CScore achieves an even better result with R = 0.8237 and RMSE = 1.0872, trained 31 on a much smaller but more relevant dataset for each target. The large data set of 32 protein-ligand complexes structural information and advances of machine learning tech-33 niques enable the data-driven approach in binding affinity prediction. CScore is capable 34 of accurate binding affinity prediction. It is also shown that CScore will perform bet-35 ter if sufficient and relevant data is presented. As there is growth of publicly available 36 structural data, further improvement of this scoring scheme can be expected. 37
doi:10.1142/s021972001100577x pmid:22144250 fatcat:uou23jrt4jc3tbxm4xxpjdkjay