Human whole-exome genotype data for Alzheimer's disease

Abstract

The heterogeneity of the whole-exome sequencing (WES) data generation methods present a challenge to a joint analysis. Here we present a bioinformatics strategy for joint-calling 20,504 WES samples collected across nine studies and sequenced using ten capture kits in fourteen sequencing centers in the Alzheimer's Disease Sequencing Project. The joint-genotype called variant-called format (VCF) file contains only positions within the union of capture kits. The VCF was then processed specifically to account for the batch effects arising from the use of different capture kits from different studies. We identified 8.2 million autosomal variants. 96.82% of the variants are high-quality, and are located in 28,579 Ensembl transcripts. 41% of the variants are intronic and 1.8% of the variants are with CADD > 30, indicating they are of high predicted pathogenicity. Here we show our new strategy can generate high-quality data from processing these diversely generated WES samples. The improved ability to combine data sequenced in different batches benefits the whole genomics research community.

Fig. 1. Summary of the VCPA-WES pipeline.

Components with “stars” are modified upon VCPA-WGS pipeline. VCPA-WES specific scripts included: CRAM metrics generation (stage1/wes_depthOfCoverage.sh), gVCF metrics generation (stage2b/wes_no_target_hc_full_bam.sh, stage2b/wes_variantEval.sh), VQSR model generation (stage 3/VQSR_snp_WES.sh, stage 3/VQSR_indel_WES.sh, stage 3/ApplyRecalibration_GATK411_SNP_WES.sh,stage3/ApplyRecalibration_GATK411_indel_WES.sh). All these can be found at https://bitbucket.org/NIAGADS/vcpa-pipeline/src/master/VCPA/.

Fig. 2. Population substructure analysis results of our dataset.

Plots from principal components analysis showing principal component (PC) 1 vs. PC2, PC2 vs. PC3, and PC1 vs. PC3 for sets of samples initially clustered on self-reported race/ethnicity (samples shown in black dots) with respect to 1kG reference populations (all other symbols). a Individual self-reporting as non-Hispanic White and clustering within 3 SD of EUR sample populations were assigned the ancestry label “Non-Hispanic White” (NHW). This plot includes 32 individuals excluded as outliers. b Individuals self-reporting as non-Hispanic Black and clustering within 3 SD of EUR and AFR sample populations or distributed between the populations were assigned the ancestry label “African American” (AFA). This plot includes 29 individuals excluded as outliers. c Individuals clustering within 3 SD of EUR and AFR sample populations and Latin American sample populations groups in the 1000 Genomes/Human Genome Diversity Project collection and between those sample population groups were assigned the ancestry label “Caribbean Hispanic” (CHI), which was also reflective of the geographic sampling of samples in the source datasets. No subjects initially classified as CHI were excluded.

Fig. 3. Jaccard similarity measure of the capture kits.

Jaccard similarity measure was calculated on all capture regions (labeled on both the x axis and y axis) at basepair level across these kits. A value of 1 (dark red) in this figure indicates that the kits are very similar to each other, while a 0 (blue) indicates the opposite.

Fig. 4. Comparison of WES CRAM quality metrics.

We compared the CRAM quality metrics across a sequencing centers (Seq_center); and b sequencing platforms (Sequencer). N for each of these 8 plots (a i to iv), b (i to iv)) all equals 20,504 subjects. Quality metrics included (i) Percentage of mapped reads, (ii) Percentage of duplicated reads, (iii) Percentage of paired reads, and iv) Quality of reads based on Q30 score. For each box plot, the center line represents the median value, the minimum of the whisker represents the 1st quantile (25th percentile), and the maximum of the whisker represents the 3rd quantile (75th percentile). Source data are provided as a Source Data file.

Fig. 5. Comparison of ×20 coverage of all the WES BAMs/CRAMs.

We compared the ×20 coverage (defined by the percentage of bps with reads or more within the sample-specific capture regions) first by a Sequencing centers, then by b Sequencers. X axis show the “×20 coverage” in percentages. N for a and b are both 20,504 subjects. For each box plot, the center line represents the median value, the minimum of the whisker represents the 1st quantile (25th percentile), and the maximum of the whisker represents the 3rd quantile (75th percentile). Source data are provided as a Source Data file.

Fig. 6. Comparison of the Ti/Tv ratio of exonic variants before and after QC.

Ti/Tv ratio is the ratio of transition (Ti) to transversion (Tv) SNPs. X axis shows the Ti/Tv ratio before QC, while the Y axis shows the Ti/Tv ratio after QC. The average Ti/Tv ratio increases from 2.53 to 3.03 after the QC process. This post-QC Ti/Tv ratio is similar to reports in previous findings. Source data are provided as a Source Data file.