By default, if a species set has finished running successfully and then you update the sample sheet to include more SRA accessions / BioSamples, snakemake will not catch this and will not do anything.
With some preliminary testing, it seems that adding -R /`snakemake --list-input-changes` to the command will work in most cases, as this will force snakemake to rerun any jobs with updated inputs as well as downstream jobs impacted by them. I have some test cases running now; there is also the option of -R `snakemake --list-params-changes` which seems to give similar but not quite identical changes.
The real question is how we should document / approach this, and if we should consider any code changes to facilitate this use case (e.g., automatically looking for changed input files somehow).
By default, if a species set has finished running successfully and then you update the sample sheet to include more SRA accessions / BioSamples, snakemake will not catch this and will not do anything.
With some preliminary testing, it seems that adding
-R /`snakemake --list-input-changes`to the command will work in most cases, as this will force snakemake to rerun any jobs with updated inputs as well as downstream jobs impacted by them. I have some test cases running now; there is also the option of-R `snakemake --list-params-changes`which seems to give similar but not quite identical changes.The real question is how we should document / approach this, and if we should consider any code changes to facilitate this use case (e.g., automatically looking for changed input files somehow).