Skip to content

Latest commit

 

History

History
108 lines (66 loc) · 5.65 KB

CHANGELOG.md

File metadata and controls

108 lines (66 loc) · 5.65 KB
Raw

v0.2.2 released - 2024-07-04

Major

  • added the --small-genomes preset. This is just an alias for -c 30 -m 200 --faster-small. This makes skani much faster when comparing hundreds of thousands of small genomes.

Minor

  • fixed a bug where skani triangle --full-matrix gave different results between STDOUT and -o (thanks to Florian Plaza Onate)
  • added a --diagonal option (suggested by Antonio Camargo) to print diagonal entries for sparse and lower-triangular distance matrices
  • added a warning to use --faster-small when comparing too many contigs (e.g. viruses, plasmids).

v0.2.1 released - 2023-10-11

More consistent support for small contigs and sequences.

Major

  • --faster-small option included in dist and triangle.

Genomes (and contigs with the --i, --ri, --qi options) with less than 20 marker k-mers are not screened according to the -s option. This makes skani for senative for small sequences, but can hamper performance on very large datasets with lots of small genomes/contigs.

This heuristic can now be disabled with the --faster-small option.

Minor

  • skani's version is now displayed properly
  • Added some error messages for degenerate cases (and more testing)
  • We found that the statically built binary can be a lot slower in certain cases. File i/o may be an issue for the binary version. A note is now added in the README.

v0.2.0 released - 2023-09-26

BREAKING

  • --learned-ani feature was buggy before and now removed.

Major

  • Major bug found: debiasing for ANI was turned off if there were > 5000 queries present in skani search and skani dist. This bug is fixed now.

Minor

  • The rust API is changing in this version. Not published to Cargo yet (waiting on DDOtten/partitions#3 to be published to crates...)

v0.1.5 released - 2023-09-01

Major

Improved "N" character support:

  • changed query-reference selection method slightly via a slight hack, using marker seeds to estimate reference length instead. This makes it so NNN characters are not counted.
  • Now seeds with "N" characters present are no longer indexed.

Minor

  • --robust now uses the learned ANI debiasing procedure by default.

v0.1.4 released - 2023-06-14

Major

  • skani triangle had a bug where if more than 5000 queries were present and --sparse or -E was not specified, the intermediate batch of 5000 queries would be written in sparse mode.
  • skani triangle -o was giving different upper triangle matrix instead of lower triangle (skani triangle > res gives lower triangle). Matrices are consistently lower triangle now.
  • Changed to lto = true for release mode. I see anywhere from a 5-10% speedup for this.

Minor

  • Changed some dependencies so no more dependencies on old crates that will deprecate.

v0.1.3 released - 2023-05-09

Major

  • Fixed a bug where memory was blowing up in dist and triangle when the marker-index was activated. For big datasets, there could be > 100 GBs of wasted memory.
  • skani now outputs intermediate results after processing each batch of 5000 queries. This will mean that outputs may no longer be deterministically ordered if there are > 5000 genomes, but you can sort the output file to get deterministic outputs, i.e skani triangle *.fa | sort -k 3 -n > sorted_skani_result.txt will guarantee deterministic output order.

Minor

  • Changed the marker index hash table population method. Used to overestimate memory usage slightly.
  • New help message for marker parameters. Turns out that for small genomes, having more markers may make filtering significantly better.
  • Added -i option to sketch so you can sketch individual records in multifastas -- does not work for search yet though, only for sketching.

v0.1.2 released - 2023-04-28.

Small fixes.

  • Added --medium pre-set, which is just -c 70. Seems to work okay for comparing fragmented genomes.
  • BREAKING: Changed --marker-index to --no-marker-index as a more sane option.
  • Added --distance option to skani triangle to output distance matrix (i.e. 100 - ANI) instead of similarity matrix.
  • Misc. help message fixes

v0.1.1 released - 2023-04-09.

Small fixes.

  • Made aligned fraction in triangle mode a full matrix by default. This is not a symmetric matrix since AF is not symmetric.
  • Misc. help message fixes

v0.1.0 released - 2023-02-07.

We added new experiments on the revised version of our preprint (Extended Data Figs 11-14). We show skani has quite good AF correlation with MUMmer, and that it works decently on simple eukaryotic MAGs, especially with the --slow option (see below).

Major

  • ANI debiasing added - skani now uses a debiasing step with a regression model trained on MAGs to give more accurate ANIs. Old version gave robust, but slightly overestimated ANIs, especially around 95-97% range. Debiasing is enabled by default, but can be turned off with --no-learned-ani.
  • More accurate aligned fraction - chaining algorithm changed to give a more accurate aligned fraction (AF) estimate. The previous version had more variance and underestimated AF for certain assemblies.

Minor

  • Small contig/genome defaults made better - should be more sensitive so that they don't get filtered by default.
  • Repetitive k-mer masking made better - smarter settings and should work better for eukaryotic genomes; shouldn't affect prokaryotic genomes much.
  • --fast and --slow mode added - alias for -c 200 and -c 30 respectively.
  • More non x86_64 builds should work - there was a bug before where skani would be dysfunctional on non x86_64 architectures. It seems to at least build on ARM64 architectures successfully now.