VariantRecalibrator produces ArrayIndexOutOfBoundsException with duplicate -an options and --rscript-file

[KevinCLydon]

Bug Report

Affected tool(s) or class(es)

VariantRecalibrator

Affected version(s)

4.3.0.0

Description

When running VariantRecalibrator, if you use -an and specify the same annotation more than once and you're also running with --rscript-file, the tool will hit an ArrayIndexOutOfBoundsException when trying to generate the R script. I think this happens because VariantRecalibrator creates an instance of VariantDataManager, which deduplicates the list of annotations to use in the data, but on line 710 in VariantRecalibrator, that list is converted into an array with length equal to the pre-deduped list of annotations, so it has more indexes than elements. Indexing into annotation lists of the data after that results in going out of bounds on the annotation lists.

See this issue in the forum: https://gatk.broadinstitute.org/hc/en-us/community/posts/13006092760859-VQSR-first-step-do-not-generate-plots