Ivan Tolstoganov: Improving genome assembly and alignment quality with synthetic long reads and multi-context seeds

Abstract

Assembly and alignment of short-read sequencing data are foundational stages in genomic analysis, as errors in these early steps can compromise downstream applications such as variant detection, transcript quantification, and metagenomic binning. Here I will present my first two PhD articles, aimed to improve the accuracy of assembly and alignment. The first paper introduces cloudSPAdes, an algorithm for synthetic long read (SLR) genome assembly, along with a universal de Bruijn graph resolution framework adaptable across SLR platforms. The second paper presents the multi-context seed construct for genome alignment, combining the benefits of short and long seeds to improve the sensitivity and specificity of sequence matching.