Glad to see rigorous discussion of how to present pop gen results. Even in standard pipelines, version info matters! Shout out the time GATK collapsed the representation of missing data with homozygous reference for "efficiency" for a couple versions:
gatk.broadinstitute.org/hc/en-us/art...
⚠️ NOTE ⚠️ This article describes behavior present in GATK versions 4.2.3.0 through 4.5.0.0, and is obsolete as of GATK 4.6.0.0, which (by overwhelming popular demand!) reverted back to the standar...
I am fully in line with the authors. The raw all-sites gVCF is the best material to share and should become the gold standard in genomics to move towards standardized and unbiased π estimates. In addition, more metadata regarding the samples and softwares used is needed, including the versions.
Thibault (Tibo) Leroy
Brault et al. propose guidelines for standardized π estimation, systematic reporting of summary statistics, and the adoption of richer metadata to ensure that population genomic data remain comparable and reusable across a wide range of systems.
🔗 doi.org/10.1093/gbe/...
#genome #evolution
