~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
* Added dependency on jobrunner and qarrayrun Python packages. These modules were formerly
part of the CFSAN SNP Pipeline and have been released as separate packages.
* Dropped support for Python 2.6 and 3.3 as some of the external packages used by the pipeline
have stopped supporting these.
* Added support for Python 3.6 and 3.7.
* VCF file generation: identify multiple alternate alleles per sample. The AD, ADF, and ADR
depths are comma-separated and ordered in the same order as the alternate alleles.
* VCF file generation: identify the alternate allele and alternate allele depth regardless of
whether the alternate allele is the consensus call.
* VCF file generation: fixed a software defect causing unreproducible VCF file results when a
sample has multiple alleles with identical depth. Older versions of the SNP Pipeline sometimes
identified different alternate alleles and different genotype on different runs with the same
datasets. In some cases, older versions of the SNP Pipeline incorrectly called the reference
allele when the snp call failed one or more filters. This change only affects VCF files. The
snp matrix file was unaffected by the software defect.
* Added the capability to control whether dense regions are filtered from each sample
independently or removed from all samples. The default behavior removes dense snp
regions from all samples, matching the behavior of previous releases. See :ref:`snp-filtering-label`
for a detailed explanation.
* Added the ``--purge`` command line option to save space by removing the intermediate output files
(the entire samples directory) upon successful completion of a run. See the
:ref:`cmd-ref-run-snp-pipeline` command reference.