v1.2.0

Highlights from previous stable release

• Locus-level TE quantification. Use -l or --locus to enable it. This will add a tilde followed by a number to TE feature name (e.g. L1PA7~123).
• Support for UMI-less libraries (e.g. smart-seq). Use with --no-umi.
• Dramatically improved memory usage efficiency and run time.
• A copy of the TE annotation generated on the fly with the -g <genome_assembly> flag will be saved in the output directory.
• Support for strand-specific count, for stranded libraries and in case knowing the orientation of transcribed TEs is relevant. Use --strandedness forward or --strandedness reverse to activate it.
• Support Python versions from 3.10 to 3.14 (NB: installation time on Python 3.14 is a few minutes longer, until Pysam's wheels update).

What's Changed

• Fix invalid escape sequence warning message by @bepoli in #19
• Set multiple verbosity level by @bepoli in #21
• Implement EM convergence with tolerance threshold of log-likelihood change by @bepoli in #20
• Add Python 3.13 test by @bepoli in #22
• SMART-seq support by @marco-cominelli01 in #27
• drop python 3.8 support by @bepoli in #28
• add test for --no-umi mode by @bepoli in #29
• Update error message for unsorted BAM in map.py by @marco-cominelli01 in #31
• Update get_ranges in misc.py by @marco-cominelli01 in #32
• Use sparse matrices in EM by @marco-cominelli01 in #35
• dependencies update by @bepoli in #36
• Support library strandedness by @bepoli in #38
• Add locus-level quantification option by @bepoli in #39
• Fix string formatting in bam file testing message by @marco-cominelli01 in #40

New Contributors

• @marco-cominelli01 made their first contribution in #27

Full Changelog: v1.1.2...v1.2.0

v1.2.0b2

Highlights

• Support for UMI-less libraries (e.g. SMART-seq). Use the --no-umi parameter to enable it.

What's Changed

• Add Python 3.13 test by @bepoli in #22
• SMART-seq support by @marco-cominelli01 in #27
• drop python 3.8 support by @bepoli in #28
• add test for --no-umi mode by @bepoli in #29

New Contributors

• @marco-cominelli01 made their first contribution in #27

Full Changelog: v1.2.0b1...v1.2.0b2

v1.2.0b1

What's Changed

• Fix invalid escape sequence warning message by @bepoli in #19
• Set multiple verbosity level by @bepoli in #21
• Converge multi-mapped UMIs estimates with tolerance threshold of log-likelihood change between EM iterations by @bepoli in #20

Full Changelog: v1.1.2...v1.2.0b1

v1.1.2

What's Changed

• Enhance compatibility with different versions of python (PR: #18)
  • Fix bug causing incompatibility with python 3.8 (a8a4549)
  • Remove support for python 3.7 (f379595)

Full Changelog: v1.1.1...v1.1.2

v1.1.1

New release to force sync with Zenodo. Equivalent to v1.1.0 release.

v1.1.0

What's Changed

• Add option "--dumpEC" to write a log table with Equivalence Classes and UMI mismatch corrections (PR: #13)
• Add options --min-bp-overlap and --min-fraction-overlap to set a minimum overlap between read and TE (PR: #14)
• Merge 1.1.0-beta.2 (PR: #16)
  • Improve the TE-UMI deduplication by using directed graphs (inspired by umi-tools directional algorithm).
  • Use networkx to handle graphs.
  • Make the EM procedure compatible with the new deduplication strategy.
  • Make EM operations on a numpy matrix to increase performance.
• Minor adjustments to docs and syntax (PR: #17)

Full Changelog: v1.0.3...v1.1.0

v1.1.0-beta.2

What's Changed

• Improve the TE-UMI deduplication by using directed graphs (inspired by umi-tools directional algorithm).
• Use networkx to handle graphs.
• Make the EM procedure compatible with the new deduplication strategy.
• Make EM operations on a numpy matrix to increase performance.

Full Changelog: v1.1.0-beta.1...v1.1.0-beta.2

IRescue version v1.1.0-beta.1

Enhancements

• Add option "--dumpEC" to write a log table with Equivalence Classes and UMI mismatch corrections (PR: #13)
• Add options --min-bp-overlap and --min-fraction-overlap to set a minimum overlap between read and TE (PR: #14)

Full Changelog: v1.0.3...v1.1.0

IRescue version 1.0.3

Bug fixes

• Fix bug causing few barcodes missing from matrix when using multiple cpus (PR: #7)
• Skip BAM lines without the chosen cell barcode tag (for increased compatibility with Cell Ranger BAM files) (PR: #11)
• Fix bug causing not deduplicating UMIs when some UMI graph configurations occur (PR: #12)

Enhancements

• Add tests (PR: #6)
• Skip multiprocessing when using 1 cpu (should make preprocessing slightly faster) (PR: #5)
• Skip chr if not in rmsk (as above) (PR: #8)
• Better error handling: provide better error description in various situations, such as those described in #1 (PR: #10)

Full Changelog: v1.0.2...v1.0.3

IRescue version 1.0.2

• bugfix: allow the presence of /1 or /2 at the end of read names (closes #1)
• bugfix: allow for more generic naming of TE features in the 4th column of the bed input file (closes #1)

Full Changelog: v1.0.1...v1.0.2