SNV Frequencies Brazil ABraOM 1k WGS Track Settings
Version: SABE-WGS-1171 Sep 2020 Assembly: Human Dec. 2013 (GRCh38/hg38)
Description
The Arquivo Brasileiro Online de
Mutações (ABraOM) provides genomic variants obtained with whole-genome sequencing
from SABE, a census-based sample of elderly individuals from São Paulo, Brazil's largest
city. The Brazilian population reflects ~500 years of admixture between Africans,
Europeans, and Native Americans. About 3% of the cohort has non-admixed Japanese ancestry
(early 20th century migration). Coverage is 38.6x. TEs, HLAs and new sequence are also available.
Data Access
The data can be explored interactively with the
Table Browser or the
Data Integrator.
For programmatic access, our REST API can be used; the
track name is abraom.
For bulk download, the VCF file can be obtained from
our download server.
The original data can also be downloaded from the ABraOM website.
Methods
For academic use only. Licensing for commercial use might be available under request and agreement.
By using this resource you agree to cite the flagship paper (Naslavsky et al. Nat Comm 2022).
Whole-genome sequencing was performed at Human Longevity Inc. using TruSeq Nano DNA HT libraries
sequenced on Illumina HiSeqX instruments with 150 bp paired-end reads targeting 30x coverage, and
reads were mapped to GRCh38 using ISIS software. Sample sex was validated by comparing CPMs of X
chromosome and male-specific Y (MSY) reads relative to autosomes, yielding the expected female
(~55,000 X CPM, <200 MSY CPM) and male (~27,500 X CPM, >550 MSY CPM) patterns. Germline SNVs
and indels were called following GATK Best Practices (GATK v3.7) via per-sample GVCFs
(HaplotypeCaller), joint genotyping (CombineGVCFs, GenotypeGVCFs), and Variant Quality Score
Recalibration (VQSR-AS); multiallelic variants were split with an in-house script, left-aligned with
BCFtools, and annotated using Annovar and custom scripts against dbSNP, 1000 Genomes, and gnomAD,
with putative loss-of-function variants identified using LOFTEE v0.3-beta irrespective of confidence
labels. Variant and genotype quality was further assessed using the in-house CEGH-Filter two-step
algorithm based on depth and allele balance, and analyses retained only GATK VQSR-AS PASS variants
and higher-confidence CEGH-Filter calls. Relatedness was assessed using KING and PC-Relate
(GENESIS), retaining a single proband per related pair and excluding one contaminated sample
(>3% by verifyBAMID), resulting in a final dataset of 1,171 unrelated individuals. Final samples
achieved mean coverages ranging from 31.3x to 64.8x, with an average of 38.65x and a median of
36.6x.
The makeDoc file documents how the source files of the varFreqs track were converted.
For some tracks, python scripts were necessary and are also available from GitHub.