4.1.0

Summary

The big news this release is primarily the update to HTSJDK 5.0.0, which brings with it to fgbio:

• 50-60% faster BAM writing on all platforms
• 30% faster BAM reading on aarch64, and modest gains (5-10%) on x86
• 40-50% faster CRAM reading and writing on all platforms
• Full CRAM 3.1 read and write support

Other Changes

• Rename duplex-filter tier locals to best/worst by @nh13 in #1156
• feat!: allow non-terminal + segment in read structures by @tfenne in #1157
• fix(codec): avoid MatchError on dovetail FR pairs with asymmetric isFrPair by @nh13 in #1155
• chore: bump sopt to 1.2.1 and scalatest to 3.2.20 by @tfenne in #1161

4.0.0

Bug Fixes

• Read structure should be mapped to all input FASTQs, not at most 3 (#1147) (4f8aa53)
• Replace broken Maven Central badge URL (#1152) (68e9561)

Documentation

• Improve cell barcode documentation across consensus calling tools (#1146) (c58bd9c)
• Update fulcrum genomics logo with light/dark theme support (#1149) (6aad37c)

Features

• Bump htsjdk to 4.2.0 and set Java 17 minimum (#1151) (d396eda)

Miscellaneous Tasks

• Bump to 3.1.3-SNAPSHOT (23b5d57)
• Update RELEASE_PROMPT.md for 3.1.2 release (9634745)
• Release 4.0.0 (51c8a8e)

3.1.2

Bug Fixes

• Grant write permission to docs job for gh-pages preview deployment (#1138) (455caa9)
• Grant pull-requests write permission to docs job for PR preview comments (#1140) (d5d5214)
• Prevent double-counting of rejected reads in CallMolecularConsensusReads stats (#1135) (#1137) (d8a1d91)

Documentation

• Specify order of clipping operations in ClipBam docs (#1143) (1cb672c)

Features

• Add a cached version of Aligner (#1060) (bf934fd)
• Log the JVM maximum runtime memory at startup (#1105) (8abba81)
• Add deterministic tie-breaking to filterToMostCommonAlignment (#1130) (7ee60c3)
• Consider a cell barcode when sorting by template coordinate (#1142) (f7474b4)

Miscellaneous Tasks

• Bump to 3.1.2-SNAPSHOT (a4096ff)
• Release 3.1.2 (56fabbf)

3.1.1

Bug Fixes

• Template.fixMateInfo() to set primary mate info before supplementary info (#1114) (da4d296)
• The consensus RX (UMI) should be called from filtered records only (#1116) (cc3b926)
• Typos in the consensus caller (#1118) (4a8644b)
• Improve the numeric stability of the consensus caller (#1120) (e2ccac9)
• Preserve input order in filterToMostCommonAlignment (#1119) (d809e1c)
• Add deterministic tie-breaking to adjacency UMI assigner (#1123) (6696d30)
• Ensure tools that use Io.assertReadable work on file descriptors too (#1113) (ebe3fe0)

Features

• Use an epsilon (2.2204460492503131E-16) when comparing log like… (#1117) (ca4bcde)
• Add --cram-ref-fasta common argument for CRAM support (#1122) (c8fbd2b)

Miscellaneous Tasks

• Bump to 3.1.1-SNAPSHOT (bba6eae)
• Release 3.1.1 (30e2245)

3.1.0

Features

• Speedup CorrectUmis and add a new --revcomp option (#1108) (329763d)
• Support cell barcodes in consensus calling tools (#1110) (7e51969)
• Support cell barcodes in ReadStructure and FastqToBam (#1103) (d7b0f04)
• Make it so FastqToBam can stream input (#1098) (8f6efe6)
• Add an entrypoint to SamSource for bytes of SAM/BAM/CRAM (#760) (2ef1130)
• Ensure file toRealPath() does not raise for file descriptors (#1106) (09e13bc)

Miscellaneous Tasks

• Bump version to 3.1.0-SNAPSHOT (1cfa5dd)
• Release 3.1.0 (033decd)

3.0.0

What's Changed

The major change in 3.0 comes in an update to our UMI-based consensus callers, and the inclusion of a new consensus caller for CODEC sequencing data (CallCodecConsensusReads). In addition to the new caller, the following updates were made:

• GroupReadsByUmi now correctly segregates read pairs with the same coordinates and UMIs but where R1 and R2 are swapped
• All consensus callers now support a --rejects option to write reads not used in consensus calling to a BAM file
  • Reads in the BAM files are also tagged with the reason they were rejected in the rr tag
• All consensus callers now support a --stats option which will output a text file with a table of statistics generated during consensus calling
• FilterConsensusReads previous --max-no-call-fraction option has been changed to --max-no-calls and is now interpreted as a fraction if less than 1.0, or as a count of no-call bases if greater than or equal to 1.0.

New Features

• Add support for collections in Metric by @nh13 in #1042
• Remove secondary and supplementary support in GroupReadsByUmi by @nh13
• Merging CODEC work into main by @tfenne in #1092
• Add support for fragment reads to TrimPrimers by @nh13 in #1067
• Add min AD and handle haploid calls for DownsampleVcf by @TedBrookings in #1079

Bug Fixes

• DownsampleVcf test was not using an expected variable by @nh13 in #1054
• AnnotateBamWithUmis did not ignore extra FASTQ records by @nh13 in #1055
• Update the CLI docs of CorrectUmis to be more functionally correct by @yfarjoun in #1082
• Remove primary name if also found in alternate names in CollectAlternateContigNames by @tfenne in #1086

Miscellaneous

• Use strict compiler flags and turn on inlining by @clintval in #1047
• Remove unused variable in AssessPhasing by @nh13 in #1052
• Do not recursively store the pr-preview directory in gh-pages by @nh13 in #1059
• Scaladoc errors to re-enable publishing by @nh13 in #1053
• Fixup sponsors formatting in main README by @clintval in #1063
• Test building with known working versions of the JDK by @clintval in #1061
• Update the skip-missing option doc across update tools by @nh13 in #1064
• Docstring for AlignmentScorer.scoreGap by @ameynert in #1071
• Update GroupReadsByUmi.scala by @cehtolonen in #1076
• Publish to maven central by @nh13 in #1083
• Allow fgbio tools to access meta information about themselves. by @nh13 in #427
• Remove most deprecated code by @nh13 in #1070
• Use a Git cliff TOML for releases by @clintval in #1062

New Contributors

• @ameynert made their first contribution in #1071
• @cehtolonen made their first contribution in #1076

Full Changelog: 2.5.21...3.0.0

2.5.21

Changelog

All notable changes to this project will be documented in this file.

[2.5.21] - 2025-04-05

⚙️ Miscellaneous Tasks

• Change version to 2.5.21-SNAPSHOT
• Debug the release artifact upload
• Release 2.5.21

2.5.20

Changelog

All notable changes to this project will be documented in this file.

[2.5.20] - 2025-04-05

⚙️ Miscellaneous Tasks

• Change version to 2.5.19-SNAPSHOT
• Need to download the JAR artifact before release
• Release 2.5.19
• Change version to 2.5.20-SNAPSHOT
• Fix rebase in github actions for gh-pages
• Release 2.5.20

2.5.17

Changelog

All notable changes to this project will be documented in this file.

[2.5.17] - 2025-04-05

🐛 Bug Fixes

• Do not provide bash args in the retry action

⚙️ Miscellaneous Tasks

• Change version to 2.5.16-SNAPSHOT
• Release 2.5.17

2.5.1

What's Changed

• ci: update actions/cache from v2 to v4 by @nh13 in #1039
• feat: FilterBam optionally writes to a rejects file by @nh13 in #1033
• chore: fix url to unit test status badge by @nh13 in #1041
• feat: add checks to ReviewConsensusVariants for FASTA fai/dict by @nh13 in #1036
• feat: improve error message in consensus calling when base qualities are missing by @nh13 in #1040
• feat: SamSource.query should handle out of range queries by @nh13 in #1028
• Added some tests to the longestDinuc to cover more edge cases by @yfarjoun in #1007
• Add an option to sort by sequence role in CollectAlternateContigNames by @nh13 in #868
• Add DownsampleVcf tool by @TedBrookings in #974
• chore: move unit test cache by @nh13 in #1045
• Add metric builder and sorter by @nh13 in #826
• ci: add preview for gh-pages to each PR by @nh13 in #1044
• ci: add sonatype release to GHA by @nh13 in #1046

New Contributors

• @TedBrookings made their first contribution in #974

Full Changelog: 2.5.0...2.5.1