Pierre Pericard

Check input test datasets are correctly downloaded before running test pipeline

2

When running the test pipeline, errors might arise because the input test datasets haven't been downloaded properly. This could lead users to think that there is a problem with the...

Ref db indexes and tmp files should not be stored in HOME directory by default

4

By default, SMR should not write in any location other than the current working directory or the directory pointed to by the `--aligned` or the `--workdir` parameter. In most bioinformatics...

Blast format help message is outdated

1

https://github.com/biocore/sortmerna/blob/4405830700ca14485aa2d4081d69cac87c429dbd/include/options.hpp#L147 In `sortmerna` help message, blast output is indicated as `(Blast - m 8 format)`. This corresponds to an outdated version of Blast, not in use anymore for more than...

Why is BLAST score different from SAM AS field score ?

2

This is something that is also present in previous SMR versions (v2.1b in particular). Alignments can be output in SAM and/or BLAST format. The Blast format requires the alignment score...

Release binaries are missing files

5

Right now, release binaries *.tar.gz files only contain the `sortmerna` binary file. At a minimum, we should include a README file, the LICENCE file, and either the rRNA databases or...

Wiki - Alignment parameter guide (table 1) is mistaking about how --best param should works

1

There are big mistakes in the alignment parameter guide from the wiki. The --best parameter do not report only the best alignment whatever value is set. The --best parameter works...

Default value of index --max_pos 1,000 or 10,000 ?

1

In SMR v4.2 help, `--max_pos` is indicated with a default value of 1,000. SMR v2.1b default value for this parameter was 10,000. Is it a simple mistake in the help...

SMR v4 now accepts as input for paired-end reads the standard two files (forward and reverse) on top of the interleaved format (forward and reverse reads in the same file)....

Ref DB index management is not up to current bioinformatics standards and is not adapted to large scale analysis in production workflows on HPC systems

1

Right now, indexing the ref DB is done automatically by SMR but in a way that is not up to current bioinformatics standards and is not adapted to large scale...

Conda best practices are to create new environments for specific tasks as much as possible. This prevents having too many packages in the base environment and ending up with incompatible...