BigBed

BigBed is a binary indexed file format used to store BED-like genomic features in a compact, searchable form. Developed by the UCSC Genome Browser project, it enables fast random access to large annotation sets by allowing queries for features that overlap a specified genomic interval without scanning the entire file. BigBed files are commonly used as an efficient replacement for plain text BED files in large-scale genome annotation collections.

Structure and data model: A BigBed file contains a binary-encoded set of BED records plus an index that maps genomic coordinates to data blocks. The records can include BED fields up to BED12, such as chrom, chromStart, chromEnd, name, score, strand, and additional optional fields. The data portion is compressed and the index supports region-based retrieval, enabling efficient extraction of features for given chromosomes and coordinate ranges.

Creation and usage: BigBed files are typically produced from BED or BED-like text files using the bedToBigBed utility, which requires a chrom.sizes file listing chromosome lengths. The resulting file, with a .bb extension, can be loaded as a custom track in the UCSC Genome Browser or accessed by software that implements the BigBed format. The Kent source package provides libraries for reading and writing BigBed files, facilitating programmatic access and integration into pipelines.

Comparison and notes: BigBed provides random access to discrete features, complementary to BigWig, which stores continuous-valued data. It is not human-readable and updates generally require converting data back to text format and re-generating the BigBed file. Support for a wide range of BED fields makes BigBed suitable for storing large, annotated genome feature sets in a scalable way.