Skip to content

Release 1.2.0

Compare
Choose a tag to compare
@nalinigans nalinigans released this 09 Dec 17:13
· 250 commits to master since this release
9e9d74c

GenomicsDB 1.2.0 jars published on Maven Central

Features in this release -

Suppress "No valid combination operation" warnings - gatk 2689 #71
Reduce memory usage when dealing with allele specific annotation fields #72
Support for mixed ploidy samples gatk 6275
Allow for using identical field names in INFO and FORMAT gatk 6158 #72

Initial GenomicsDB 64-bit Integer support for INFO fields. This is especially required to compute annotation fields for DP, MQ fields when the number of samples is large (~100K) and exceed the 32-bit Integer space.
Limitations and caveats:
(a) Only INFO fields are currently supported for 64-bit ints.
(b) Note that the BCF2 format does not support 64-bit ints for reading/writing. However, the VCF format, being a text format, has no such limitations and will work fine.
(c) If the input VCF has fields with 64-bit ints, the vid information must specify that it's a 64-bit int field ("type": "int64" in the vid json).
(d) Existing fields which are 32-bit ints in the input VCFs (such as DP, MQ etc.) will work fine. The output of the sum will be a 64-bit value. If the output value fits within 32-bits, everything works exactly as before. If not, the resulting VCF record can only be output in the VCF format (BCF not supported).