Skip to content

Latest commit

 

History

History
95 lines (78 loc) · 3.75 KB

CHANGES.md

File metadata and controls

95 lines (78 loc) · 3.75 KB
Raw

v0.2.3 (dev)

  • for snps add 0's to DHGT if none overlap

v0.2.2

  • update for newer htslib
  • better support for manta (and likely other callers) (#38 thanks to @am8265 for reporting)

v0.2.1

  • fix for start of SV less than 99th percentile of insert size distribution (#28, #29, thanks @roryk)

v0.1.5

  • support for more ALT types (e.g DUP:TANDEM, DEL:ME, etc)
  • flush stdout properly (#25)

v0.1.4

  • output DHSP for BND elements on same chrom
  • fix for #22: allow GQ of type Integer or Float

v0.1.3

  • add DUPHOLD_SAMPLE_NAME environment variable as a way to set the sample name. (#19)
  • re-vamp SNP stuff so that only DHGT is reported. It is an Integer with 5 columns for numbers of 0. hom-ref 1. het 2. hom-alt 3. unknown 4. low-quality snp variants within each event.

v0.1.2

  • remove PCRE dependency.
  • reduce logging output.
  • support un-indexed VCFs without contig definitions.
  • use only PASS SNVs from --snps
  • set DHFFC flank distance to 1000 default (was 5000) and allow setting via DUPHOLD_FLANK env var and only use non-zero values in DHFFC flank. These 2 changes should improve DHFFC for samples with sparse coverage (e.g. for genomes without good assemblies). (Thanks @JoWhi for reporting and providing a test-case)

v0.1.1

  • fix bug when later chroms had larger values (see: 9f88400 and thanks @raul-w for reporting and providing a test-case).
  • adjust calculation of global and per-gc bin coverage to use only non-zero bases. this improves the ratio when comparing to the depth inside of events (which will still count zero-coverage bases) in chromosomes with sparse coverage. (thanks @raul-w for suggesting).

v0.1.0

  • reduce memory usage in discordant calculation.

v0.0.9

  • drop DHD in favor of DHFFC for (F)lank fold change.
  • make changes to support long, single-end reads.
  • fix bug in hts-nim that made duphold grab the wrong sample info from a SNP BCF (not vcf)

v0.0.8

  • skip duphold calculation for distant or interchromosomal BND's
  • don't decode CRAM qname. This makes duphold use ~15% less CPU for CRAM
  • allow annotating SVs with SNP calls as a additional validation of each event.

v0.0.7

  • remove DHBZ for the z-score. this was less useful than the fold-change values.
  • greatly improve DHD. the cutoffs for this are now data-driven so that we expect a low false-positive rate of ends that are called to have a rapid change in depth. this requires an extra pass over each chromosome that will add about 90 seconds of runtime per (human or similar-sized) genome.
  • DHBFC is now based on the median instead of mean so it is less susceptible to outliers and off-center calls
  • fix bug that in rare cases resulted in exiting with "how"
  • also report DHSP for number of supporting read-pairs for the event. This is usually redundant with SVTyper info, but might be more accurate in some cases.

v0.0.6

  • duphold will now annotate the space between BND's that occur within 20MB on the same chromosome. Sometimes events that are obvious deletions by their change in coverage will be called as a pair of BNDs. This will help prioritize those that are accompanied by a depth change.

v0.0.5

  • fix bug that occurred with SVs near either end of chromosome (#2). thanks Brad for reporting and providing a test-case.

v0.0.4

  • small bug-fixes
  • duphold now uses about half as much memory
  • add -d/--drop flag which will drop all samples from a VCF except the one matching the sample in --bam. this simplifies per-sample parallelization followed by merge.

v0.0.3

  • fix bug when annotatign multi-sample vcf (#2)
  • get sample name from bam read group info. NOTE that this changes the command line parameters by removing the --sample argument.