Skip to content

Latest commit

 

History

History
232 lines (199 loc) · 11.9 KB

CHANGES.md

File metadata and controls

232 lines (199 loc) · 11.9 KB
Raw

v0.3.1

v0.3.0

  • bugfix for 0.2.8 regression where SNPs present in gnotate file would be instead annotated as missing #149. This happened when annotating a chromosome after a chromosome present in the query set that was not in the gnotate file

v0.2.9

  • bugfix for 0.2.8 regression where indels present in gnotate file would be instead annotated as missing #149

v0.2.8

  • [internal] use same zip library for make-gnotate as gnotate. this should improve speed for reading many small chromosomes.
  • don't quit on CSQ/ANN/BCSQ fields that don't have enough information to parse (#122)
  • fix segregating_dominant_x

v0.2.7

  • [tsv] fix for empty VCFs

v0.2.6

  • allow specifying QUAL and ID to -i argument to slivar tsv (#104 #105)

v0.2.5

  • it was previously not possible to adjust the order of impacts with SLIVAR_IMPACTFUL_ORDER. This is now fixed. (#97)

v0.2.4

  • fix long-standing bug (#27) that would cause sigsegv in some cases

v0.2.3

  • add feature_fusion to impact order list (#92)
  • [compound-hets] write summary even if no variants found

v0.2.2

  • fix bug with '.' in ALT field (caused message about incorrect number of alts in some cases. thanks Batsal for reporting)
  • slivar compound-hets: fix bug with parents specified in ped file but absent from VCF (#79)
  • slivar compound-hets: add "intergenic_region" to list of impacts that are skipped by default
  • add SLIVAR_NO_REPORT_ALL to prevent reporting variants for familys with no affected samples
  • fix bug when creating huge zip files with make-gnotate (#86)
  • NOTE: change default min depth in slivar-functions.js to 6. (was 0)
  • don't hard-code tmp directory to /tmp (use $TMPDIR)
  • [compound-hets] dont fail if no usable variants were found, just issue warning. this can happen for small chroms or regions when slivar was parallelized.

v0.2.1

  • add INFO.highest_impact_order so user can do things like:

    --info "INFO.highest_impact_order == ImpactOrder.missense"

    to get only missense mutations. or

    --info "INFO.highest_impact_order <= ImpactOrder.missense"

    to get variants with an impact of missense or higher. (note that lower values have higher impact)

  • tsv: bugfix!! previously, slivar tsv used a < 0 check where it should hvae used <= 0 so candidates with a parental sample at index 0 in the samples of the vcf would have missing information for depth, GQ, and AB. (#78 . Thanks @amwenger for finding the problem and its source).

  • expr: always annotate variant and INFO even if no sample (or trio,group,fam) expressions are given so one can use slivar just to get the 'impactful' annotation.

  • remove duplicates from default-order.txt

v0.2.0

  • clear values from expressions for family, group, and sample. was previously only done for trios. so if you relied on something like: 
    --group-expr "expr1: ..." \
--group-expr "expr2:('expr1' in INFO) && ..."\
    then the expr2 would still see expr1 in INFO even if the preceding expression did not pass but it has passed for a previous variant. Thanks Amelia Wallace for finding and reporting a simple test-case.
  • add /opt/slivar/slivar-functions.js to docker image (#75)

v0.1.13

  • fix bug with VCFs with >256 info fields (#74). Thanks @liserjrqlxue for making a great test-case.

v0.1.12

  • slivar expr: handle ploidy > 2 (updated hts-nim) (#55 thanks to @markw3lsh for reporting and providing a test-case)
  • slivar: optionally support strings in format (sample) fields. default is to ignore these for performance reasons, but users can set the environment variable SLIVAR_FORMAT_STRINGS to anything to force slivar to populate the javascript objects with any strings in the VCF. e.g. SLIVAR_FORMAT_STRINGS=1 slivar expr ... and use unset SLIVAR_FORMAT_STRINGS to remove. (#37)
  • slivar compound-hets: fix bug with 2 affect sibs sharing same compound-het previously only 1 sib would be reported

v0.1.11

  • gnotate: ~10-20% speed improvement by inlining cmp in binarysearch
  • slivar expr: fix counts in summary table when some families in the cohort have no affected samples
  • slivar tsv: improve readability when 1 or both parents are missing
  • slivar tsv: fix comma separation with 1 or both parents missing (thanks @brwnj for reporting)
  • slivar ddc: when looking at all chromosomes, skip X, Y
  • slivar expr: add --exclude option which takes a bed file of exclude regions
  • slivar compound-hets: default will now output compound-het pairs that include synonymous variants
  • pslivar: fix tsv output from pslivar (which allows running slivar in parallel). more info here
  • add INFO.genic boolean to complement INFO.impactful. By default this includes all impacts included by impactful along with synonymous and other exonic, but non-protein-altering variants (but does not include intronic. (more info here

v0.1.10

  • [slivar ddc] fix bug when selecting sample filters.
  • [slivar ddc] UI fix: remove button when [x] is clicked.

v0.1.9

v0.1.8

  • fix bug when subsetting vcfs (ped file contained subset of samples in vcf). Thanks Matt for reporting.

v0.1.7

  • slivar make-gnotate will error with warning if field is not a float or int
  • compound-het provides a way to ignore some impacts (intron, non_coding, etc) and sets sane defaults. this removes most candidates from WGS which were predominantly pairs of intronic or non-coding variants.
  • compound-het: remove -f flag, this is now discovered by default
  • only samples with parents in the vcf (not just ped) are counted as trios
  • expose sample.phenotype attribute which is a string taken directly from the pedigree file
  • given an error on an INFO expression, don't bail, just report and continue
  • fix for VCFs without index AND without contigs in header (#44)
  • support for file sizes > 4.2GB and support for gnotate files with chr prefix annotating query files without
  • fix gnotate bug where some long alleles were not annotated

v0.1.6

  • add INFO.impactful (boolean) if CSQ/BCSQ/ANN is present. this value will be true of any of those annotation are high enough
  • the impactful flag is automaticaly added to the output VCF for any variant passing through slivar and meeting the criteria (https://github.com/brentp/slivar/wiki/impactful)
  • better checks for length of AD field
  • fix for empty groups (#38)
  • add --family-expr for family expressions like: fam.every(function(s) { return s.het == s.affected && s.hom_ref == !s.affected && s.GQ > 5 })
  • slivar tsv now outputs a sortable column for highest-impact. it makes use of a default list of impact orderings from: https://uswest.ensembl.org/info/genome/variation/prediction/predicted_data.html and supplemented with any type seen in bcftools or snpEff.
  • bug fix with gnotate (-g) when annotating a file with a empty chromsome followed by a non-empty (e.g. if chr16_random is mixed in with canonical chromosomes)

v0.1.5

  • !!MAJOR: fix bug with hom_alt alias for alts == 2. Since v0.1.3, hom_alt would not get set to false for samples, variants after being set once.
  • update wiki to simplify rare disease analysis.
  • tune javascript in my.js and rare-disease expressions so that many analyses will run better than twice as fast.
  • [expr] add --skip-non-variable which further improves speed by not evaluating javascript for a trio if all members are hom-ref or unknown (but can the same variant will still be evlauated for trios with at least 1 sample that has a variant at that site.
  • [compound-het] major speed increase for large cohorts.
  • [compound-het] remove -i/--index option as this is now discovered from the header for the CSQ/BCSQ field.

v0.1.4

  • better error messages on bad VCF
  • [tsv] add --csq-column to slivar tsv to allow extracting extra CSQ fields
  • general usability improvements
  • [compound-het]: support singleton (kid only) or duo's (kid and 1 parent) to compound het. this will give many false positives because the variants can't be phased, but the number can still be quite small given sufficient filtering on population allele frequency. requires --allow-non-trios argument

v0.1.3

  • add new sample attributes hom_ref, het, hom_alt, unknown which are synonyms of alts == 0, alts == 1, alts == 2, alts == -1 respectively.
  • remove slivar gnotate sub-command. the same functionality can be had from slivar expr with -g and --info.

v0.1.2

  • new sub-command slivar tsv to output a tab-separated value file from a filtered VCF for final processing this command also allows adding columns based on the gene the variant is in, such as description of gene function, pLI, etc.
  • fix bug where slivar expr would segfault if labels were re-used (now quits with error message)
  • better help in slivar make-gnotate
  • improve and document output format for duo-del https://github.com/brentp/slivar/wiki/finding-deletions-in-parent-child-duos
  • add a prelude function with hasSample(INFO, key, sample) to see if sample has been added to info field by previous filter.
  • allow outputting summary table to file with SLIVAR_SUMMARY_FILE environment variable
  • fix for make-gnotate with multiple files (still requires first file to contain all chromosomes)
  • allow more kinds of pedigree files (#30)
  • change format of slivar compound het field: adds an id that uniquely identifies the pair of variants.

v0.1.1

v0.1.0

  • allow accessing INFO via variant.INFO as well as INFO
  • better error message on missing gnotate file
  • add --sample-expr to slivar expr sub-command to allow applying an expression to each sample: `--sample-expr "hi_quality:sample.DP > 10 && sample.GQ > 20"
  • expose samples in javascript in $S object, e.g. $S["sampleABC"].DP > 10 && $S["sampleXYZ"].DP > 10
  • better checking of matching between ped and vcf => don't fail on missing kids

v0.0.9

  • fix bug without --pass-only. slivar always behaved as if --pass-only was used.
  • [compound-hets] fix for multi-sample VCFs with --sample-field

v0.0.8

  • use random file names so concurrent slivar processes don't clobber files.
  • more informative error on bad js expression
  • fix for empty groups (#20)
  • fix bug when later expressions depended on previous ones.

v0.0.7