v2.3.6.1

What's Changed

No code changes from 2.3.6.0 but rebuilding due to typo/errors in 2.3.6.0 tag

Full Changelog: v2.3.5.0...v2.3.6.1

v2.3.6.0

What's Changed

• refine_assembly: tolerate whitespace in fasta headers by @dpark01 in #49

Full Changelog: v2.3.5.0...v.2.3.6.0

v2.3.5.0

What's Changed

• bump viral-core to 2.3.5 by @dpark01 in #48

Full Changelog: v2.3.4.0...v2.3.5.0

v2.3.4.0

What's Changed

• bump viral-core 2.3.2 to 2.3.4 by @dpark01 in #47

Full Changelog: v2.3.2.0...v2.3.4.0

v2.3.2.0

What's Changed

Build updates & upstream viral-core dependency update.

• update coveralls action; bump actions python to 3.10 by @tomkinsc in #45
• docbuild fixes, viral-core bump by @dpark01 in #46

Full Changelog: v2.3.1.4...v2.3.2.0

v2.3.1.4

What's Changed

• skani updates by @dpark01 in #44
  • First, this changes all skani invocations to use -m 15 -c 10 as the new default parameters for the marker k-mer compression factor (-m) and the k-mer subsampling rate (-c). This is down from our previous default values of -m 30 -c 20 which is in turn lower than the skani-recommended settings for viruses of -m 200 -c 30 which is lower than the skani bacterial-inspired default values of -m 1000 -c 125. The new default values were determined through empirical testing of diverse viral taxa, and -m 15 -c 10 succeeds in clustering (finding non-zero ANI between) all of rhinovirus & enterovirus together as well as all of Lassa virus (our old default values only partially succeeded on these taxa).
  • Second, this PR reformats skani TSV output to be written out in sorted order of descending ANI * Total_bases_covered (instead of the default sort order of descending ANI). This prevents our reference selection code from favoring reference genomes with higher ANI for short stretches of sequence (which happened with some rhino/enteros) and now favors longer matches of high identity. This specific metric (the product of ANI and match length) is inspired by ReferenceSeeker (publication, GitHub). Unit tests added to confirm proper sort order.

Full Changelog: v2.3.1.3...v2.3.1.4

v2.3.1.3

What's Changed

• expose common parameters for skani at argparse level by @dpark01 in #43

Full Changelog: v2.3.1.2...v2.3.1.3

v2.3.1.2

What's Changed

• bugfix: make skani dist tolerate empty input by @dpark01 in #42

Full Changelog: v2.3.1.1...v2.3.1.2

v2.3.1.1

What's Changed

• bugfix: assemble.skani_contigs_to_refs filtered file was missing header by @dpark01 in #40

Full Changelog: v2.3.1.0...v2.3.1.1

v2.3.1.0

What's Changed

• skani-based reference selection tools by @dpark01 in #39
• update viral-core 2.3.0 -> 2.3.1 by @tomkinsc in #38
• addressing documentation build failures; testing ReadTheDocs build check for PRs by @tomkinsc in #31

Full Changelog: v2.3.0.0...v2.3.1.0