Figure 1

Algorithmic workflow of ALFALFA. ALFALFA follows a canonical seed-and-extend workflow for mapping reads onto a reference genome. The reference genome is indexed by an enhanced sparse suffix array (a) to enable quick retrieval of variable-length seeds called super-maximal exact matches between a read and the reference genome (b). Seeds are then grouped into non-overlapping clusters that mark candidate genomic regions for read alignment (c). Handling of candidate regions is prioritized by agglomerate base pair coverage of the seeds. The final extend phase samples seeds from candidate regions to form collinear chains that are bridged using banded dynamic programming (d). All of these steps strive to make optimal reuse of seeds in order to avoid superfluous computations. Background image used with permission from Walter Obermayer.