Main description:

This thesis addresses practical data quality and variant detection issues regarding research on next-generation DNA sequencing. It develops for the first time a hidden Markov model (HMM) to describe the error pattern and data generation of next-generation DNA sequencers. Further, it proposes using a haplotype-based method, which employs the HMM and re-alignment to suppress the interference of sequencing errors and incorrect alignments, in order to improve the detection accuracy of SNPs and InDels. Lastly, the thesis sheds new light on the interpretation and application of error modeling for next-generation DNA sequencing.

Contents:

Introduction.- Statistical Model for Next Generation Sequencing.- PyroHMMsnp: A Re-Alignment Based SNP Caller.- PyroHMMvar: An Alignment-Graph Based InDel Caller.- Conclusion and Perspective.