Zinc-binding protein action site prediction method based on integrated learning in non-equilibrium mode

Abstract

The invention discloses a zinc-binding protein action site prediction method based on integrated learning in a non-equilibrium mode. According to characteristics of the zinc-binding protein action site, protein source data is pre-processed; non-equilibrium of the zinc-binding protein site is equilibrated through the random down-sampling technology to obtain several sub-equilibrium data sets; for the sub-equilibrium data sets, distinguishable protein biochemical characteristics are selected for characteristic representation to form characteristic vectors; the characteristic vectors are respectively taken as the input of a base classifier support vector machine to calculate the sample weight, a probability neural network model based on the sample weight is constructed, and lastly, the base classifier support vector machine and the probability neural network model based on the sample weight are integrated to obtain a prediction model; the prediction model is utilized to identify the zinc-binding protein action site of a target sample.

Description

technical field [0001] The invention relates to a zinc-binding protein action site prediction method based on integrated learning in an unbalanced mode, which aims at identifying the zinc-binding protein action site by using an integrated learning classification model in an unbalanced classification mode, and belongs to proteomics and computer science the cross field. Background technique [0002] With the completion of the Human Genome Project, life science has entered the post-genome era, and the protein expressed by genes has become one of the important research topics in the fields of life science and natural science. Protein is the basic organic matter that constitutes cells, the material basis of life, and plays a decisive role in the process of biological life. However, this decisive role is not simply determined by a single protein. In most cases, the protein needs to interact with other proteins or ligands to complete specific biological functions. [0003] In cel...

AI Technical Summary

Problems solved by technology

At present, the existing prediction methods use data mining and other methods to establish classification models, treat the two types of samples equally, and do not take into account the imbalance of data, resulting in low accuracy in the prediction of zinc-binding protein interaction sites

Images