Some Models And Algorithms Of Group Genomics

With the advance in biotechnology, vast amounts of genomic data to population genomics has brought unprecedented opportunities and challenges. Population genomics is increasingly dependent on computer automation process and the corresponding efficient statistical methods. In this thesis, the authors present statistical models and computer algorithms as well as corresponding benckmarks to several practical problems encountered in population genomics. The "single nucleotide polymorphism genotyping algorithm" and "copy number variation genotyping algorithm" allow us call biological significant genotype data from raw signals generated by experimental equipments. In the "pairwise haplotype" and "simple copy number variation haplotype" problems of the study, the authors derived the haplotype frequency estimation and the corresponding variances with mathematical analytic expression. For the "chromosome-scale haplotype reconstruction" and " haplotype construction containing copy number variation" issue, the authors proposed fast and accurate algorithms and discusses the use in practics. Finally, the author attempts to build a new statistic in the genome-wide search for the positive natural selection experienced, and found that part of the genes in the UV-adaptive system, olfactory system and the neural stem cell development system showed obvious signs of positive selection. These problems have a certain representation in population genomics and their software implementation are of practical value. Several solutions to the problems have reached the world advanced level.