v1.1.6: Update version.py

UMI-tools output is now deterministic with --random-seed

Many users have had issues with making UMI-tools deterministic, which previously relied upon both --random-seed and the enivornmental variable PYTHONHASHSEED being set. From v1.1.6 only --random seed is required.

Please note that in some cases the implemented solution may make the output from v.1.1.6 different to previous versions, even if --random-seed is set to the same value. The differences will be very slight and the different outputs represent equally sensible UMI grouping/deduplication since they relate only to how ties are broken.

Thank you @TyberiusPrime, @christianbioinf and others for their suggestions for how to remove the dependency on PYTHONHASHSEED for deterministic output.

New features

• umi_tools is now deterministic when using --random-seed - @TomSmithCGAT in #550
• Option to extract barcode from read2 only - @TomSmithCGAT in #630
• Adds support for python 3.12 - @IanSudbery in #657

Bugfix

• Avoids switching matplotlib backend - @sshen8 in #640
• count_tab now correctly reads UMI and cell barcodes - @eachanjohnson in #654
• count_tab now writes out strings not bytes - @eachanjohnson in #654
• Installation with < python 3 prevented - @IanSudbery in #644

Documentation

• FAQ entry regarding identification of possible duplicates reads/pairs - @TomSmithCGAT in #631
• Improved docs regarding chimeric/unmapped/unpaired read pairs - @TomSmithCGAT in #629

Other

• Add issue templates - @TomSmithCGAT in #632
• Update testing suite to pytest - @eachanjohnson in #655

New Contributors

• @sshen8 made their first contribution in #640
• @eachanjohnson made their first contribution in #654

Full Changelog: 1.1.5...v1.1.6