Method for predicting affinity of HLA type I molecule and polypeptide

Abstract

The invention discloses a method for predicting the affinity of HLA I type molecules and polypeptides. According to the method, the advantages of various algorithms are integrated systematically, anda data set obtained by collection of an IEDB database and literature research is integrated with a data set generated through an experiment into a database; numeric type conversion is carried out on peptide fragment corresponding to each HLA type in the data set; and a final training data set is obtained. Deep research on different coding modes of different learner combinations and analysis on data characteristics are carried out; according to the method, a mixed model of multiple algorithms is integrated, the characteristics of the polypeptide sequence are learned, independence of single typing on training of a machine learning algorithm is guaranteed, and the prediction complementarity of HLA typing on a deep learning algorithm is also guaranteed, so that the prediction of the affinity of the polypeptide and specific HLA molecules is realized; and further, the combination of a plurality of machine learning tools is utilized to accurately predict the neoantigen in the tumor of the patient.

Description

technical field [0001] The invention relates to the field of biological information, in particular to a method for predicting the affinity between HLA type I molecules and polypeptides. Background technique [0002] Tumor immunotherapy is an emerging tumor treatment method. Its core technology is to target tumor neoantigens, stimulate the patient's immune system, differentiate and proliferate immune effector cells that specifically target tumor cells, and precisely act on tumors carrying neoantigen targets. cell. Therefore, tumor immunotherapy can theoretically eliminate all tumor cells carrying specific antigens without harming normal cells. In contrast, traditional tumor treatment, whether it is surgical resection of diseased tissue, or taking chemical drugs to kill cells that proliferate too fast, or radiation irradiation, will cause varying degrees of damage to the normal cells of the patient, bringing many problems to the patient. pain. [0003] The process of the im...

AI Technical Summary

Problems solved by technology

The sources of peptide sequences with affinity to HLA mainly include binding experiments and mass spectrometry analysis. There are many literatures and academic works that have reported these experimental data, but there is currently no database that comprehensively summarizes these data; the market needs a method that combines various The advantages of the algorithm have been systematically integrated to generate a prediction model that can improve the classification effect, high stability, and high prediction accuracy

Images