Accelerating Pangenome Analysis on De Bruijn Graphs<\/p>\n
Z\u00fclal Bing\u00f6l
\nPhD Student
\n(Supervisor: Prof.Dr.Can Alkan )<\/p>\n
Computer Engineering Department
\nBilkent University
\nAbstract: Most genomics analyses begin with or involve a comparison with the reference genome of the same species, which serves as a template for the newly sequenced sample of an individual. Since the early days of genomics analyses, a single reference genome, that contains an individual\u2019s genomic information, has been used. However, using the single reference creates reference bias since it does not include all of the variations showing the genetic diversity within the population. Because a single reference does not represent different individuals, the formation of the reference to incorporate a collection of genomes as a pangenome is becoming more valid and preferred in sequence analysis pipelines. Pangenomes consist of many genomes; therefore, representing them in an efficient format that stores a large amount of data compactly while enabling effective access to variance regions on the genomes is crucial. From this perspective, graphs offer a flexible storage solution for pangenomes by facilitating multiple connections within the data. Once the reference is constructed, the analyses usually start by read mapping to find the positions of the newly sequenced sample\u2019s reads on the reference genome. Regarding execution time, read mapping is the bottleneck of the following steps because of its computationally-costly alignment sub-procedure. With increased and potentially more complex data, mapping the reads to the pangenome graph reference is even more daunting. Therefore, improving the execution time of read mapping to pangenome graphs is crucial for downstream analyses. In this work, our goal is to find solutions for optimizing the pangenome graphs for less storage and developing efficient algorithms that eventually lead to fast and accurate mapping. This thesis, therefore, provides guidelines for an accelerator framework for pangenome analysis.<\/p>\n
DATE: September 17, Thursday @ 14:00
\nPLACE: Mithat \u00c7oruh Amphiteathre<\/p>\n","protected":false},"excerpt":{"rendered":"
Accelerating Pangenome Analysis on De Bruijn Graphs Z\u00fclal Bing\u00f6l PhD Student (Supervisor: Prof.Dr.Can Alkan ) Computer Engineering Department Bilkent University Abstract: Most genomics analyses begin with or involve a comparison with the reference genome of the same species, which serves as a template for the newly sequenced sample of an [\u2026] <\/p>\n