Nicola Rizzo: Indexable sequence graphs for pangenomics

Abstract: Computational pangenomics is an emergent subfield of biological sequence analysis and its goal is to replace linear reference genomes with diverse, unbiased, and efficient collections of genomes. A sequence graph is an expressive and compact representation for a collection of high-quality genomes, but it presents theoretical and practical challenges around the basic tasks of exact pattern matching and sequence alignment. In this talk, I will review a few key results on indexing linear sequences with the Burrows–Wheeler transform and present different attempts at generalizations to graph indexes: Wheeler graphs, elastic founder graphs, and the generalized compressed suffix array (GSCA2).