Reconstruction Of Three-dimensional Structure Of Single Cell Chromosome Based On Optimization Method

In biology,it is generally believed that the structure of biomacromolecules affects their functions.The three-dimensional structure of chromosome plays an important role in the regulation of gene expression and the occurrence of genetic disease.With the generation and application of chromosome conformation capture technology,people began to use Hi-C technology to obtain chromosome interaction data,and using these data to reconstruct the three-dimensional structure of chromosome has become an important research direction in the field of bioinformatics.In this paper,the distance constraint model for reconstructing the three-dimensional structure of chromosomes is studied.Firstly,the Hi-C data is processed by the recursive graph method to get the preliminary distance of chromosome segments? Due to the single cell Hi-C data,there are too many missing data.We combine the shortest distance method and interpolation method to get a relatively complete frequency matrix of chromosome fragment interaction,further get the distance matrix of chromosome fragment,and then use the classical multidimensional scale analysis method to get the preliminary three-dimensional structure of chromosome.Because there are errors in the processing of Hi-C data and the completion of low rank matrix,the genetic algorithm is used to optimize the above distance model.In the optimization process,there is a "premature" phenomenon of genetic algorithm,which will make the distance with smaller fitness function value eliminated completely in several iterations,and the final result is not comprehensive.The optimization method combined with genetic algorithm optimizes the distance constraint model,combines the m-reception criteria with selection,crossover and mutation,and eliminates the disadvantages brought by the "precocity" of genetic algorithm.In this paper,through the combination of the shortest distance method and the interpolation method,the missing information in the contact frequency matrix with low rank can be recovered better.Furthermore,the genetic algorithm is used to simulate the natural selection process,select the individuals with higher fitness function,and eliminate the errors brought by data processing.However,in the long-term selection process,some individuals with lower fitness function value are eliminated,resulting in the final Results a part of the information of the contact sites was missing.The evolutionary algorithm combined simulated annealing algorithm and genetic algorithm was used to make the final fitness function of the model the highest.