Anticancer peptide recognition method based on fusion of random forests and related vector machine

Abstract

The invention relates to an anticancer peptide recognition method based on r fusion of random forests and a related vector machine. The method comprises the following specific steps: carrying out feature extraction on the composition of amino acids, determining the average percentage of each amino acid in ACP and non-ACP, and determining the sequence features of ACP; constructing an RVM model by determining priori distribution and posteriori distribution and performing iterative computation; training the RVM model, calculating a posteriori mean value and a posteriori variance, and predicting atraining sample; and subjecting a given sample to feature sampling by adopting an RRVM algorithm using the given sample as a sample for RVM modeling, and when the feature of a new peptide chain is input, carrying out prediction by using the RVM model and judging whether the input new peptide chain is ACP or non-ACP. The method is superior to a traditional simple method adopted by most of researchers to identify ACP at present, solves the problem of low identification precision, and fully provides technical support for biological experiments to reduce the cost and improve efficiency.

Description

technical field [0001] The invention relates to the technical field of anticancer peptide identification, and relates to an anticancer peptide identification method based on the fusion of random forest and correlation vector machine. Background technique [0002] Humans have developed many technologies to control and kill cancer: Traditional methods such as radiotherapy, targeted therapy and chemotherapy can suppress cancer to a certain extent, while expensive costs and side effects of treatment and cancer cells are not effective against current anticancer chemotherapy drugs Drug resistance is an unavoidable defect of these treatment options. [0003] In 1972, Boman discovered the primary structure of the antimicrobial peptide of hyaluronic acid. Later, many researchers found that antimicrobial peptides have antitumor activity. They then named the antimicrobial peptide an anticancer peptide (ACP). ACP has many advantages, such as high specificity, low production cost, hig...

AI Technical Summary

Problems solved by technology

Antimicrobial peptides (AMPs) can be used to obtain ACP, and many cationic AMPs can destroy bacteria but not normal cells, and are cytotoxic to a variety of cancer cells

Although the mechanism of ACP is still not fully understood, the development of natural ACP and artificially designed peptides is still an important way to fight cancer

[0004] However, experimental techniques are expensive and time-consuming methods to find ACPs, thus identification of ACPs by computational methods is a necessary means to solve the problem

However, most of the current researchers use traditional and simple methods such as support vector machine (SVM) and artificial neural network (ANN) to identify ACP, which leads to low recognition accuracy and cannot fully provide technical support for biological experiments to reduce costs. ,Improve efficiency

Images