Molecular recognition feature function prediction method based on ensemble learning

Abstract

The invention discloses a molecular recognition feature function prediction method based on ensemble learning, and mainly solves the problem that an existing molecular recognition feature predictor cannot further divide molecular recognition feature functions. According to the scheme, the method comprises the following steps: downloading inherent disordered protein data and functional annotations thereof, dividing training data and test data, performing feature representation on a protein sequence, and designing a residue tag of the protein sequence; selecting a single-input binary association strategy machine learning model; training different machine learning models by using the training data; integrating training results of different machine learning models by using an integration strategy to collect a prediction model; and inputting to-be-researched protein sequence data into the prediction model, and outputting a molecular recognition feature function prediction result of the protein. The method is simple in experimental process, low in resource consumption, low in cost and high in reliability of prediction results, can be used for predicting molecular recognition features in protein sequences, and provides reference for drug target acting positions.

Description

[0001] The invention belongs to the technical field of bioinformatics, and in particular relates to a method for predicting the function of molecular recognition features, which can be used to predict the function of molecular recognition features in protein sequences and provide references for drug target action positions. Background technique [0002] Molecular recognition features refer to regions in proteins that contain inherently disordered regions of between 10 and 70 residues that transform from disordered regions to ordered regions after binding to their partners. The partners include carbohydrates, ions, lipids, nucleic acids, proteins and small molecules. The functions of molecular recognition features include molecular recognition assembler (molecular recognition assembler), molecular recognition scavenger (molecular recognition scavenger), molecular recognition effector (molecular recognition effector), molecular recognition display sites (molecular recognition dis...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to provide a protein molecular recognition feature function prediction method based on integrated learning to solve the problems of complex experimental procedures, large resource consumption and high cost in the prior art.

Images