Development of algorithms for the analysis of duplex sequencing data

Abstract

Duplex sequencing detects ultra-rare mutations by tagging DNA molecules with double-stranded tags. This method creates single-stranded consensus sequences (SSCS) from the reads, which then form duplex consensus sequences (DCS) and are then aligned to the reference genome to call mutations. During this process, a large amount of sequencing data is lost. Therefore, we have developed new algorithms, that give insight in the sequencing data which helps to improve the reads/SSCS/DCS ratios. In addition, a graphical representation of the sequencing data was implemented. The first part of the thesis is focusing on the distribution of sizes of read families. Second, a detailed analysis of the tags is shown by calculating their Hamming distances which can identify sequencing or PCR errors from true molecules. In addition, we can detect artificial produced chimeric reads during PCR. The fourth part includes the application of our algorithms on shorter tag lengths and on only those tags which are involved in the formation of DCSs. Finally, we investigated different sources of read loss during data analysis.