Fishing in read collections: Memory efficient indexing for sequence assembly

Abstract

In this paper, we present a memory efficient index for storing a large set of DNA sequencing reads. The index allows us to quickly retrieve the set of reads containing a certain query k-mer. Instead of the usual approach of treating each read as a separate string, we take an advantage of significant overlap between reads and compress the data by aligning the reads to an approximate superstring constructed specifically for this purpose in combination with several succint data structures.