BGEN

This is a history of releases of the BGEN reference implementation.

v1.1.6 (March 13 2020)

• Update bgenix to use PRAGMA synchronous=OFF; when indexing, addressing issue #42. Thanks to Jesse Stacey for reporting this and investigating the fix. In my tests on GPFS filesystem this speeds up indexing considerably, with similar speedups seen for other filesystems.

v1.1.5 (January 1st 2020)

• Output variant IDs separated by semicolon, not command in bgenix -vcf.
• Doc updates

v1.1.4 (August 7th 2018)

• bgen library bugfix for BGEN v1.1 files and sample subsets

v1.1.3 (May 1st 2018)

• bgenix -vcf: further speed up output (by about ~20X; d0b786b)

v1.1.2 (Apr 26 2018)

• bgenix: support NFSv3 filesystems by avoiding locking #13 (154d089d5ddf)
• bgenix -vcf: bugfix for lookup-table based implementation on some platforms #16 (0b911770f33c)

v1.1.1 (Mar 23 2018)

• bgenix: lookup-table based implementation of bgenix -vcf for 8-bit encoded files
• additional file reading optimisations
• implement functional tests / continuous integration tests (see pipelines)
• full rbgen source package can now be constructed with ./waf build_rbgen

v1.1.0 (Feb 5th 2018)

• Performance improvements for bgen file reading
• rbgen: improvements to build process
• rbgen: bgen.load(): add rsids argument to restrict to specified rsids
• rbgen: bgen.load(): add samples argument to restrict to specified samples
• bgenix: be more sensible about behaviour when there's an existing partially created index (.bgi.tmp) file
• implement example compute_expected_dosage. This computes expected dosage for biallelic variants using either floating-point arithmetic or a lookup table, enabling performance comparisons.
• Implement Robot Test Framework-based functional testing

v1.0.2 (Jan 25th 2018)

• bgenix: fix VCF output metadata

v1.0.1 (Jul 27th 2017)

• bgenix: fix for VCF output of ID fields

• THis release