Skip to content

Latest commit

 

History

History
201 lines (127 loc) · 5.99 KB

CHANGELOG.md

File metadata and controls

201 lines (127 loc) · 5.99 KB

8.6.0

  • Allow users to keep original soft-mask with the new option --keepOriginalMask

8.5.0

8.4.0

  • Paint poly-N regions in dotplots. Theses usually are contig boundaries.

8.3.0

  • Experimental protein mode triggered when the seed is PSEUDO. Implementation may change in the future.

8.2.1

  • Ignore failures of last-train instead of crashing. The failures are usually caused by the lack of similarity between target and query genomes, and it is better to let the other alignments proceed to the end.

8.1.0

  • The suffix of the trained parameter files is changed from .par to .train.

  • New experimental one to many mode (--o2m) which may be useful when all possible ways of alignment are needed. Note that the name of the output files may be changed in the future.

8.0.0

  • Use default -D value (total input length) for last-train. Keep -D1e9 for lastal.

  • Upgrade LAST to version 1541. This introduces seed-specific defaults for the maximum repeat unit length when the target genome is soft-masked by lastdb.

  • Rename the --seeding_scheme option to just --seed, which is shorter and easier to remember.

  • Rename the --skip_m2m (default true) option to just --m2m (default false).

  • Rename the files and options to clearly indicate m2m, m2o and o2o.

7.1.0

  • Output a text-formatted trace file to profile resource usage.

  • Reduce the number of CPUs of last-split tasks to 2.

  • Update to LAST 1522 to allow for RY128 seeds.

7.0.0

  • --skip_m2m now defaults to true.

  • Add a --one_to_one_only option to prevent copying the lastal alignment to the results folder, thus saving disk space.

  • Add a --lastal_extra_args option to pass lastal arguments that are not recognised by last-train.

  • Change the suffix of the parameter file from par to 00.par for better sorting of the file names.

  • Stop providing a copy of the LAST index in the results folder.

  • Index both strands to speed up computation (at the expense of memory usage).

  • Add a --last_split_mismap option and revert the default to 1e-5.

  • Update LAST to version 1519 and windowmasker to version 2.15.0.

  • Default to soft-mask lowercased letters (option -c of lastdb), and make the postmask step optional.

  • Replace the -E0.05 option (“Maximum expected alignments per square giga”) with -D1e9 (“Report alignments that are expected by chance at most once per LENGTH query letters”) to match the tutorials closer. Both options should have similar effects, but -D is easier to explain.

6.1.0

  • New --read_align option to utilise the pipeline for mapping query reads to a target genome.

6.0.0

  • New --skip_m2m to skip the generation of the many-to-many alignment, which consumes a large amount of time and disk space.

  • Change default -m value of last-split to the default (-m1 at the moment) and add a new option --last_split_args to allow setting other values (such as -m1e-5 that was used previously).

  • New --targetName option to include target genome names in the output files.

5.2.2

  • Guess index file name by searching for prj files and selecting the shortest base name. The previous method failed when the indexed genome was large enough to cause the generation of multiple prj (or des) files. Version 5.2.1 attempted to solve the problem but failed.

5.2.0

  • New --dotplot_options option to modify the dot plots. New default sort and orientation of the query genome (to match the alignment to the target genome). query genome sequence names are now written horizontally.

  • In the README's examples, reversed the role of the target and query sequences for better demonstrating the new dotplot defaults.

5.1.0

  • New --with_windowmasker option to soft-mask the genome with the windowmasker tool of the BLAST suite.

5.0.0

  • Move postmask step at the end of the workflow, so that last-split has more information.
  • Use the new --reverse option of last-split so that the use of maf-swap (and the files it generates) can be avoided.
  • Pass fMAF+ to the fist call of last-split and -m1e-5 to the second call, as in the upstream cookbook.
  • Update LAST to version 1250.
  • Use YASS as default seed and advertise RY32 in the README.

4.2.0

  • New options --skip_dotplot_1, _2, and _3 to skip computationally expensive and not always so useful plots.

4.1.1

  • Correct version number in nextflow.config and brush up documentation.

4.1.0

  • Optionally pass a single alignment parameter file with a new --lastal_params option. Doing so skips last-train.
  • Re-implement the correction of 4.0.0 in a way that complies with nf-core, using a join operation.

4.0.0

  • Important bug fix ensuring that the right trained parameter set is used with the right genome. (2c05b2de2da69864020fc4203f15d2fa14350d9c)

3.1.1

  • Update LAST modules to the version accepted in nf-core.

3.1.0

  • New --query option that saves the effort of creating a sample sheet when there is only one query genome.

3.0.0

  • Force the score matrix to be symmetric (pass --revsym to last-train).
  • Allow passing common arguments to last-train and lastal with the --lastal_args option, defaulting to -E0.05 -C2.

2.0.1

  • Correct a bug that caused index names to not be detected properly when seeding schemes such as MAM8 are used.

2.0.0

  • New --seeding_scheme that defaults to NEAR. In previous versions the lastdb command did not receive a parameter and defaulted to YASS.

1.1.1

  • Solve a channel bug that prevented processing more than one sample.

1.1.0

  • Add dotplots.

1.0.1

  • Set computation resources for process labels.

1.0.0

  • Initial version.