2f44ab096235810d5d621b7356bc64fbebe82494
lrnassar
  Wed May 13 13:02:57 2026 -0700
varFreqs: fix numeric-claim discrepancies on indigenomes.html and sgdpFreq.html.

indigenomes.html: clarify that the deployed VCF is the public release subset
(18,016,257 records) of the larger Jain 2021 callset (55.8M variants), and
note that the public release is sites-only with a VRT variant-type INFO field
and no AC/AF.

sgdpFreq.html: update Methods to reflect the deployed file (44,756,737 SNV
records, 601,775 multiallelic decomposed); drop the "34.4M SNPs + 2.1M indels"
claim; clarify that the Mallick 2016 FermiKit indel callset is not carried in
this track. refs #36642

diff --git src/hg/makeDb/trackDb/human/sgdpFreq.html src/hg/makeDb/trackDb/human/sgdpFreq.html
index 2493a74358b..c1ff40f6849 100644
--- src/hg/makeDb/trackDb/human/sgdpFreq.html
+++ src/hg/makeDb/trackDb/human/sgdpFreq.html
@@ -27,38 +27,41 @@
 track name is <em>sgdpFreq</em>.
 For bulk download, the VCF file can be obtained from
 <a href="http://hgdownload.soe.ucsc.edu/gbdb/hg38/varFreqs/" target="_blank">our download server</a>.
 </p>
 
 <p>The original source VCFs are available from
 <a href="https://sharehost.hms.harvard.edu/genetics/reich_lab/sgdp/vcf_variants/"
 target="_blank">https://sharehost.hms.harvard.edu/genetics/reich_lab/sgdp/vcf_variants/</a>.
 </p>
 
 <h2>Methods</h2>
 <p>
 High-coverage whole-genome sequencing of 300 individuals (279 publicly available) from 142
 diverse populations was performed on Illumina instruments using PCR-free library preparation at
 an average depth of 43x. Reads were aligned to the hs37d5 reference (GRCh37 with decoy
-sequences) using BWA-MEM 0.7.12. SNP and indel genotyping was performed using GATK
-HaplotypeCaller with joint genotyping across all samples. An independent indel callset was
-generated using FermiKit for improved sensitivity at complex variants. The final dataset
-contains 34.4 million SNPs and 2.1 million short indels.
+sequences) using BWA-MEM 0.7.12. SNP genotyping was performed using GATK
+HaplotypeCaller with joint genotyping across all samples. (The Mallick 2016 release also
+includes an independent indel callset generated with FermiKit; indels are not carried in
+this track.)
 </p>
 <p>
-The VCFs were merged with bcftools and lifted to hg38 with CrossMap. At UCSC, genotypes were
-stripped to produce a sites-only frequency VCF retaining the existing AC, AF, and AN INFO fields.
+The per-sample VCFs were merged with bcftools and lifted to hg38 with CrossMap. At UCSC,
+genotypes were stripped to produce a sites-only frequency VCF retaining the AC, AF, and AN
+INFO fields. The deployed file contains 44,756,737 SNV records (601,775 of which represent
+multiallelic sites split into separate biallelic records). Indels from the source callset
+are not included.
 We provide documentation that indicates how all source files were converted in the <a href="https://github.com/ucscGenomeBrowser/kent/blob/master/src/hg/makeDb/doc/hg38/varFreqs.txt" target="_blank">makeDoc file</a> of the track.
 Python scripts are also available from <a href="https://github.com/ucscGenomeBrowser/kent/blob/master/src/hg/makeDb/scripts/varFreqs" target="_blank">GitHub</a>.
 </p>
 
 <h2>Credits</h2>
 <p>
 This project was funded by the Simons Foundation. Thanks to David Reich and Swapan
 Mallick for help with importing the data.
 </p>
 
 <h2>References</h2>
 <p>
 Mallick S, Li H, Lipson M, Mathieson I, Gymrek M, Racimo F, Zhao M, Chennagiri N, Nordenfelt S,
 Tandon A <em>et al</em>.
 <a href="https://doi.org/10.1038/nature18964" target="_blank">