New Tools
Added markdup, a tool to locate and tag duplicate reads in a BAM or SAM file, where duplicate reads are defined as originating from a single fragment of DNA.
-
Major performance improvements.
-
Supports read lengths of up to 500 base pairs.
-
Updated to GATK 4.3.
-
Supports read lengths of up to 500 base pairs.
-
Adds support for the
--htvc-bam-output
option. -
--batch
is deprecated.
-
Updated to GATK 4.3.
-
Supports read lengths of up to 500 base pairs.
-
Adds support for the
--run-partition
option.
-
Updated to GATK 4.3.
-
Clarified error messages.
Improvements spanning multiple tools
-
Added more error checking when writing BAM files using
--gpuwrite
. -
Improved performance for BWA alignment, particularly for reads > 250 bases in length.
-
Added GPUDirect Storage (GDS) support for fq2bam (BWA-MEM + GATK) and associated pipelines.
-
Added a low memory mode for
--gpuwrite
(affects sorting).