Unexpected behavior when reading multiple gVCFs with tabix indices and same base name

[williambrandler]

When reading multiple gVCFs that have tabix indices and the same base name. It seems that the index becomes corrupted when the second gVCF is read and slicing based on chromosome no longer works as expected.

The workaround is to simply delete the tabix index or change the basename, so probably something to do with how the index is stored.

here's an example of paths that cause the issue,
gvcf_path_1 = "/mnt/path/to/test_data/version_1/NA12878_markdup_realigned_recalibrated_Haplotyper.g.vcf.gz"
gvcf_path_2 = "/mnt/path/to/test_data/version_2/NA12878_markdup_realigned_recalibrated_Haplotyper.g.vcf.gz"