Software for Recognition of DNA Secondary Structures Associated with Epigenetic Factors by Machine Learning Methods

The purpose of this final qualification work is developing research software(Software) the software for the recognition of patterns allowing to reveal correlation between secondary structures of DNA and epigenetic factors, methods of machine learning for a clustering "without teacher". Developed Software it will be used by erudite biologists as the tool in a research of secondary structures of DNA of people. The best algorithm of classification was chosen and built-in in Software for the solution of problems of classification. In this final qualification work for the analysis of data sets and the qualifier of templates Pandas, NumPy, LightGBM, Sklearn, Scikit-learn and Matplotlib libraries were used. Possibilities of the program: — Finding of areas of correlation between secondary structures of DNA and their epigenetic factors and result conclusion in a graphic form. — Possibility of creation by the user of a clustering. — Creation of model of classification and finding of patterns. — Creation of the file of the report. At the first stage of work preparation of data has been made for the further analysis. At the second stage has been written the cross-platform application by means of which regularities have been found and the correlation between them is counted. Highly correlating sites were included into a basis of the training selection in which 84 parameters for each gene in the site are counted. Also the user interface for convenience of work with the program has been written. All computing code is written on Python, and the user interface has been created by means of Tkinter library. The volume of the report on final qualification work, not including applications, is – NN pages. The number of tables in work – 59, illustration – 43, the used sources – 32.