Protein function prediction method combining multi-task learning and self-attention mechanism

Abstract

The invention discloses a protein function prediction method combining multi-task learning and a self-attention mechanism, and the method comprises the following steps: predicting a task according toa molecular function type MF, a biological process type BP, and a cell module type CC; constructing a protein function prediction system model based on multi-task learning and a self-attention mechanism; obtaining a sample data set, extracting feature information of protein sequences in the sample data set, and establishing a training set and a test set; preprocessing the training set, inputting the preprocessed training set into a protein function prediction system model, and training the protein function prediction system model; and preprocessing the test set, inputting the preprocessed testset into the trained protein function prediction system model, and carrying out protein function prediction. According to the method, prediction of three ontology theories is regarded as three prediction tasks, prediction is performed by establishing a protein function prediction system model based on multi-task learning and a self-attention mechanism, and the accuracy of protein function prediction is improved.

Description

technical field [0001] The invention relates to the field of protein function prediction, and more specifically, relates to a protein function prediction method combining multi-task learning and self-attention mechanism. Background technique [0002] Protein function prediction is a very important task in the field of biology, and plays a key role in the development of new drugs and pathological understanding. In the early days, the functional annotation of proteins was mainly carried out through in vivo or in vitro experiments, but its time-consuming and expensive characteristics could not adapt to the development speed of current high-throughput sequencing technology. Therefore, calculation-based methods have gradually become the focus due to their low cost and fast speed. research direction. [0003] Protein function can be annotated by Gene Ontology (Gene Ontology), which contains more than 40,000 functional items, which can be divided into three categories: molecular f...

AI Technical Summary

Problems solved by technology

[0007] In order to solve the lack of use of multi-task learning methods in the prior art and the difficulty in accurately predicting protein functions, the present invention provides a protein function prediction method combining multi-task learning and self-attention mechanism

Images