GATK VariantsToTable: QUAL="." in input becomes QUAL="-10.0" in output

[friedav]

Bug Report

Affected tool(s) or class(es)

gatk VariantsToTable

Affected version(s)

The Genome Analysis Toolkit (GATK) v4.5.0.0

Description

Hi,

when running the VariantsToTable tool on a VCF file in which no genotype data are included and no QUAL information is given (i.e. .), the QUAL column in the output table contains -10.0.

In an attempt to understand this unexpected behavior, I traced it back to the getPhredScaledQual() function in
https://github.com/samtools/htsjdk/blob/master/src/main/java/htsjdk/variant/variantcontext/CommonInfo.java,
where getLog10PError() * -10) + 0.0 is returning -10.0, thus getLog10PError() seems to be returning 1.

This is some example input and output:

$ # columns in input VCF: no genotype data
$ zcat variants_chr2.vcf.gz  | grep '#CHROM'
#CHROM  POS     ID      REF     ALT     QUAL    FILTER  INFO

$ # in input VCF, QUAL is "."
$ zcat variants_chr2.vcf.gz  | grep -v '##' | cut -f1-8 | head
#CHROM  POS     ID      REF     ALT     QUAL    FILTER  INFO
2       10797   rs28888107      C       T       .       PASS    IMPUTED;AF=0.0127291;MAF=0.0127291;AVG_CS=0.998805;R2=0.925789
2       11336   rs113656530     C       G       .       PASS    IMPUTED;AF=0.0282261;MAF=0.0282261;AVG_CS=0.997855;R2=0.940762
2       11357   rs111385029     A       G       .       PASS    IMPUTED;AF=0.0282287;MAF=0.0282287;AVG_CS=0.997853;R2=0.940673
2       11486   rs73138514      A       G       .       PASS    IMPUTED;AF=0.0283085;MAF=0.0283085;AVG_CS=0.997773;R2=0.938685
2       11594   rs114792740     G       T       .       PASS    IMPUTED;AF=0.0183994;MAF=0.0183994;AVG_CS=0.99201;R2=0.654262
2       11607   rs73138516      T       C       .       PASS    IMPUTED;AF=0.0282278;MAF=0.0282278;AVG_CS=0.997854;R2=0.940712
2       11834   rs73910134      A       G       .       PASS    IMPUTED;AF=0.0282808;MAF=0.0282808;AVG_CS=0.9978;R2=0.938963
2       11842   rs13390778      C       G       .       PASS    IMPUTED;AF=0.0282808;MAF=0.0282808;AVG_CS=0.9978;R2=0.938963
2       11944   rs10172629      C       T       .       PASS    IMPUTED;AF=0.0124383;MAF=0.0124383;AVG_CS=0.998878;R2=0.928775

$ gatk VariantsToTable -V variants_chr2.vcf.gz  -O variants_chr2.tsv

$ # in input table, QUAL is "-10.0"
$ head variants_chr2.tsv
CHROM   POS     ID      REF     ALT     QUAL    FILTER  AF      MAF     AVG_CS  R2      ER2     IMPUTED TYPED
2       10797   rs28888107      C       T       -10.0   PASS    0.0127291       0.0127291       0.998805        0.925789        NA      true    NA
2       11336   rs113656530     C       G       -10.0   PASS    0.0282261       0.0282261       0.997855        0.940762        NA      true    NA
2       11357   rs111385029     A       G       -10.0   PASS    0.0282287       0.0282287       0.997853        0.940673        NA      true    NA
2       11486   rs73138514      A       G       -10.0   PASS    0.0283085       0.0283085       0.997773        0.938685        NA      true    NA
2       11594   rs114792740     G       T       -10.0   PASS    0.0183994       0.0183994       0.99201 0.654262        NA      true    NA
2       11607   rs73138516      T       C       -10.0   PASS    0.0282278       0.0282278       0.997854        0.940712        NA      true    NA
2       11834   rs73910134      A       G       -10.0   PASS    0.0282808       0.0282808       0.9978  0.938963        NA      true    NA
2       11842   rs13390778      C       G       -10.0   PASS    0.0282808       0.0282808       0.9978  0.938963        NA      true    NA
2       11944   rs10172629      C       T       -10.0   PASS    0.0124383       0.0124383       0.998878        0.928775        NA      true    NA

Expected behavior

If QUAL = . in input and no genotype data present, QUAL = . in output

Actual behavior

If QUAL = . in input, QUAL = -10.0 in output

In case this is the expected behavior and not a bug, I would appreciate some insights in why that is.

Thanks,
Friederike