Why it matters

BWT is foundation of bzip2 + BWA. Understanding shapes compression + bioinfo.

Advertisement

The architecture

Sort all cyclic rotations.

Last column = BWT.

Invertible with LF-mapping.

BWT structureCyclic rotationssort lexicographicallyLast columnBWT stringInvert via LF maprecover originalFM-index built on BWT for space-efficient searchable text; bzip2 uses for compression
BWT.
Advertisement

How it works end to end

Construction: SA-based O(n).

Inverse: LF-mapping.

FM-index: search on BWT in low space.