v1.9.0

New:

• add --unpairedOutFastqto taxon_filter.py trim_trimmomatic: two optional parameters for specifying paths to fastq files containing Trimmomatic’s unpaired reads. If the parameters are not specified, the unpaired reads are written to temp and discarded.
• optional intrahost params are now exposed in the Snakemake config.yaml file:
  • intrahost.py vphaser_one_sample --minReadsEach and --maxBias
  • intrahost.py merge_to_vcf --naive_filter
• The plot_coverage function now accepts --plotOnlyNonDuplicates that only includes reads not marked with a 1024 bitmask.
• The align_and_plot_coverage function now accepts --excludeDuplicates, which performs Picard’s MarkDuplicates on the aligned bam, and then calls plot_coverage with --plotOnlyNonDuplicates. Other parameters are added to change the JVMmemory and pass opts to Picard.
• Novoalign parameter strings for the Snakemake rules refine_assembly_1, refine_assembly_2, and map_reads_to_self have been exposed in the config.yaml file
• Downsampling before de novo assembly is now performed in bam space via Picard's DownsampleSam. This is faster and allows unpaired reads to be included in cases where the paired read count is below the Trinity input size threshold.

Fixed:

• Fix Trimmomatic usage bug: explicitly use Phred-33 quality scores. See PR #409 for more information.
• Memory limit raised for lastal rule (3->8G), and Snakemake merge_one_per_sample rule (7->10G) to suit larger project sizes
• For Broad users, the Snakemake rule filter_to_taxon is set to use the long UGER queue

Updated:

• Picard updated, v1.126 -> v1.141