samtools/bcftools 1.19

on GitHub

Download the source code here: bcftools-1.19.tar.bz2.(The "Source code" downloads are generated by GitHub and are incomplete as they don't bundle HTSlib and are missing some generated files.)

Changes affecting the whole of bcftools, or multiple commands:

• Filtering expressions can be given a file with list of strings to match, this was previously possible only for the ID column. For example

   ID=@file            .. selects lines with ID present in the file
   INFO/TAG=@file.txt  .. selects lines where TAG has a string value listed in the file
   INFO/TAG!=@file.txt .. TAG must not have a string value listed in the file

• Allow to query REF,ALT columns directly, for example -e 'REF="N"'

Changes affecting specific commands:

• bcftools annotate

  • Fix bcftools annotate --mark-sites, VCF sites overlapping regions in a BED file were not annotated (#1989)

  • Add flexibility to FILTER column transfers and allow transfers within the same file, across files, and in combination. For examples see https://samtools.github.io/bcftools/howtos/annotate.html#transfer_filter_to_info

• bcftools call

  • Output MIN_DP rather than MinDP in gVCF mode

  • New -*, --keep-unseen-allele option to output the unobserved allele <*>, intended for gVCF.

• bcftools head

  • New -s, --samples option to include the #CHROM header line with samples.

• bcftools gtcheck

  • Add output options -o, --output and -O, --output-type

  • Add filtering options -i, --include and -e, --exclude

  • Rename the short option -e, --error-probability from lower case to upper case -E, --error-probability

  • Changes to the output format, replace the DC section with DCv2:

    • adds a new column for the number of matching genotypes

    • The --error-probability is newly interpreted as the probability of erroneous allele rather than genotype. In other words, the calculation of the discordance score now considers the probability of genotyping error to be different for HOM and HET genotypes, i.e. P(0/1|dsg=0) > P(1/1|dsg=0).

    • fixes in HWE score calculation plus output average HWE score rather than absolute HWE score

    • better description of fields

• bcftools merge

  • Add -m modifiers to suppress the output of the unseen allele <*> or <NON_REF> at variant sites (e.g. -m both,*) or all sites (e.g. -m both,**)

• bcftools mpileup

  • Output MIN_DP rather than MinDP in gVCF mode

• bcftools norm

  • Add the number of joined lines to the summary output, for example
    Lines total/split/joined/realigned/skipped: 6/0/3/0/0

  • Allow combining -m and -a with --old-rec-tag (#2020)

  • Symbolic <DEL> alleles caused norm to expand REF to the full length of the deletion. This was not intended and problematic for long deletions, the REF allele should list one base only (#2029)

• bcftools query

  • Add new -N, --disable-automatic-newline option for pre-1.18 query formatting behavior when newline would not be added when missing

  • Make the automatic addition of the newline character in a more predictable way and, when missing, always put it at the end of the expression. In version 1.18 it could be added at the end of the expression (for per-site expressions) or inside the square brackets (for per-sample expressions). The new behavior is:

    • if the formatting expression contains a newline character, do nothing
    • if there is no newline character and -N, --disable-automatic-newline is given, do nothing
    • if there is no newline character and -N is not given, insert newline at the end of the expression
      See #1969 for details

  • Add new -F, --print-filtered option to output a default string for samples that would otherwise be filtered by -i/-e expressions.

  • Include sample name in the output header with -H whenever it makes sense (#1992)

• bcftools +spit-vep

  • Fix on the fly filtering involving numeric subfields, e.g. -i 'MAX_AF<0.001' (#2039)

  • Interpret default column type names (--columns-types) as entire strings, rather than substrings to avoid unexpected spurious matches (i.e. internally add ^ and $ to all field names)

• bcftools +trio-dnm2

  • Do not flag paternal genotyping errors as de novo mutations. Specifically, when father's chrX genotype is 0/1 and mother's 0/0, 0/1 in the child will not be marked as DNM.

• bcftools view

  • Add new -A, --trim-unseen-allele option to remove the unseen allele <*> or <NON_REF> at variant sites (-A) or all sites (-AA)