Compound target protein binding prediction method based on multi-deep learning model consensus

Abstract

The invention discloses a compound target protein binding prediction method based on multi-deep learning model consensus. The method comprises the following steps: acquiring compound target protein binding / non-binding data; extracting a plurality of different compound target protein binding / non-binding data sets for training, testing and verifying; respectively constructing and training a plurality of compound target protein binding / non-binding deep learning models, extracting compound target protein binding / non-binding characteristics, and obtaining a plurality of final compound target protein binding / non-binding deep learning models; predicting to obtain a plurality of groups of compound target protein binding / non-binding relationships through the plurality of compound target protein binding / non-binding deep learning models in the previous step; integrating the obtained multiple groups of binding relationship results to obtain a consensus binding relationship, and integrating the obtained multiple groups of non-binding relationship results to obtain a consensus non-binding relationship. The method has the characteristics of low false positive rate and high accuracy, and is suitable for popularization and application.

Description

technical field [0001] The invention belongs to the technical field of pharmaceutical research and development, and relates to a compound target protein binding prediction method based on consensus of multiple deep learning models. Background technique [0002] Use the existing mainstream deep learning models (fully connected neural network model, convolutional neural network model, recurrent neural network model, encoder-decoder network and deep belief network) to extract the characteristics of the binding relationship of the compound target protein, so that Predicting the binding relationship of new compounds to target proteins has important practical significance for the discovery / development of new drugs and the study of the mechanism of action of traditional Chinese medicine. However, although these models can achieve high training accuracy (>90%), the high prediction false positive rate (>90%) prevents the further application of deep learning models in this field...

AI Technical Summary

Problems solved by technology

[0003] Existing techniques exploit compound-target protein binding relationships (see figure 1 ) cheminformatics software, all manual, many parameters need to be adjusted, the success rate is low, and there is a problem of high false positive

Deng, N et al (J.Phys.Chem.B 2015,119,976-988.) and Nataraj S et al (Biophys Rev (2017)9:91–102.DOI10.1007 / s12551-016-0247-1) respectively From the two aspects of compound target protein binding free energy calculation (free energy calculation) and compound target protein binding software review, it is discussed that there are many parameters in the compound target protein binding prediction, the success rate is low, and there are high false positives.

Images