bedtools intersect -a snATAC_peaks.bed -b credible_sets.bed -wa -wb -u    cellranger-atac count --reference GRCh38 --fastqs runs/ --id muscle_01
macs2 callpeak -t muscle.nuclei.bam -f BAMPE -g hs --nomodel --shift -100    STAR --runThreadN 16 --genomeDir GRCh38 --readFilesIn R1.fq.gz R2.fq.gz
plink2 --pfile cohort --glm hide-covar --covar covariates.tsv --out eqtl    nextflow run main.nf -profile slurm,singularity -resume
Rscript -e 'susieR::susie_rss(z, R, n = 402, L = 10)'    snakemake --profile slurm --use-singularity --jobs 400
Rscript -e 'coloc::coloc.abf(gwas, qtl, p12 = 1e-5)'    sbatch --array=1-402%40 --mem=32G --time=24:00:00 map_qtl.sh
QTLtools cis --vcf geno.vcf.gz --bed expression.bed.gz --permute 1000    tabix -p vcf imputed.chr10.dose.vcf.gz
samtools view -b -q 30 -F 1804 aln.bam chr10:112990000-113010000    bcftools norm -m -any -f GRCh38.fa | bcftools view -q 0.01:minor
cellranger-atac count --reference GRCh38 --fastqs runs/ --id muscle_01    liftOver credible_sets.hg19.bed hg19ToHg38.over.chain.gz out.bed unmapped.bed
STAR --runThreadN 16 --genomeDir GRCh38 --readFilesIn R1.fq.gz R2.fq.gz    ldsc.py --h2-cts muscle.sumstats.gz --ref-ld-chr baseline. --w-ld-chr weights.
nextflow run main.nf -profile slurm,singularity -resume    macs2 bdgcmp -t treat.bdg -c control.bdg -m FE -o fold_enrichment.bdg
snakemake --profile slurm --use-singularity --jobs 400    gzip -dc gwas.sumstats.gz | sort -k1,1 -k2,2n | bgzip > sorted.gz
sbatch --array=1-402%40 --mem=32G --time=24:00:00 map_qtl.sh    bedtools intersect -a snATAC_peaks.bed -b credible_sets.bed -wa -wb -u
tabix -p vcf imputed.chr10.dose.vcf.gz    macs2 callpeak -t muscle.nuclei.bam -f BAMPE -g hs --nomodel --shift -100
bcftools norm -m -any -f GRCh38.fa | bcftools view -q 0.01:minor    plink2 --pfile cohort --glm hide-covar --covar covariates.tsv --out eqtl
liftOver credible_sets.hg19.bed hg19ToHg38.over.chain.gz out.bed unmapped.bed    Rscript -e 'susieR::susie_rss(z, R, n = 402, L = 10)'
ldsc.py --h2-cts muscle.sumstats.gz --ref-ld-chr baseline. --w-ld-chr weights.    Rscript -e 'coloc::coloc.abf(gwas, qtl, p12 = 1e-5)'
macs2 bdgcmp -t treat.bdg -c control.bdg -m FE -o fold_enrichment.bdg    QTLtools cis --vcf geno.vcf.gz --bed expression.bed.gz --permute 1000
gzip -dc gwas.sumstats.gz | sort -k1,1 -k2,2n | bgzip > sorted.gz    samtools view -b -q 30 -F 1804 aln.bam chr10:112990000-113010000
bedtools intersect -a snATAC_peaks.bed -b credible_sets.bed -wa -wb -u    cellranger-atac count --reference GRCh38 --fastqs runs/ --id muscle_01
macs2 callpeak -t muscle.nuclei.bam -f BAMPE -g hs --nomodel --shift -100    STAR --runThreadN 16 --genomeDir GRCh38 --readFilesIn R1.fq.gz R2.fq.gz
plink2 --pfile cohort --glm hide-covar --covar covariates.tsv --out eqtl    nextflow run main.nf -profile slurm,singularity -resume
Rscript -e 'susieR::susie_rss(z, R, n = 402, L = 10)'    snakemake --profile slurm --use-singularity --jobs 400
Rscript -e 'coloc::coloc.abf(gwas, qtl, p12 = 1e-5)'    sbatch --array=1-402%40 --mem=32G --time=24:00:00 map_qtl.sh
QTLtools cis --vcf geno.vcf.gz --bed expression.bed.gz --permute 1000    tabix -p vcf imputed.chr10.dose.vcf.gz
samtools view -b -q 30 -F 1804 aln.bam chr10:112990000-113010000    bcftools norm -m -any -f GRCh38.fa | bcftools view -q 0.01:minor
cellranger-atac count --reference GRCh38 --fastqs runs/ --id muscle_01    liftOver credible_sets.hg19.bed hg19ToHg38.over.chain.gz out.bed unmapped.bed
STAR --runThreadN 16 --genomeDir GRCh38 --readFilesIn R1.fq.gz R2.fq.gz    ldsc.py --h2-cts muscle.sumstats.gz --ref-ld-chr baseline. --w-ld-chr weights.
nextflow run main.nf -profile slurm,singularity -resume    macs2 bdgcmp -t treat.bdg -c control.bdg -m FE -o fold_enrichment.bdg
snakemake --profile slurm --use-singularity --jobs 400    gzip -dc gwas.sumstats.gz | sort -k1,1 -k2,2n | bgzip > sorted.gz
sbatch --array=1-402%40 --mem=32G --time=24:00:00 map_qtl.sh    bedtools intersect -a snATAC_peaks.bed -b credible_sets.bed -wa -wb -u
tabix -p vcf imputed.chr10.dose.vcf.gz    macs2 callpeak -t muscle.nuclei.bam -f BAMPE -g hs --nomodel --shift -100
bcftools norm -m -any -f GRCh38.fa | bcftools view -q 0.01:minor    plink2 --pfile cohort --glm hide-covar --covar covariates.tsv --out eqtl
liftOver credible_sets.hg19.bed hg19ToHg38.over.chain.gz out.bed unmapped.bed    Rscript -e 'susieR::susie_rss(z, R, n = 402, L = 10)'
ldsc.py --h2-cts muscle.sumstats.gz --ref-ld-chr baseline. --w-ld-chr weights.    Rscript -e 'coloc::coloc.abf(gwas, qtl, p12 = 1e-5)'
macs2 bdgcmp -t treat.bdg -c control.bdg -m FE -o fold_enrichment.bdg    QTLtools cis --vcf geno.vcf.gz --bed expression.bed.gz --permute 1000
gzip -dc gwas.sumstats.gz | sort -k1,1 -k2,2n | bgzip > sorted.gz    samtools view -b -q 30 -F 1804 aln.bam chr10:112990000-113010000
bedtools intersect -a snATAC_peaks.bed -b credible_sets.bed -wa -wb -u    cellranger-atac count --reference GRCh38 --fastqs runs/ --id muscle_01
macs2 callpeak -t muscle.nuclei.bam -f BAMPE -g hs --nomodel --shift -100    STAR --runThreadN 16 --genomeDir GRCh38 --readFilesIn R1.fq.gz R2.fq.gz
plink2 --pfile cohort --glm hide-covar --covar covariates.tsv --out eqtl    nextflow run main.nf -profile slurm,singularity -resume
Rscript -e 'susieR::susie_rss(z, R, n = 402, L = 10)'    snakemake --profile slurm --use-singularity --jobs 400
Rscript -e 'coloc::coloc.abf(gwas, qtl, p12 = 1e-5)'    sbatch --array=1-402%40 --mem=32G --time=24:00:00 map_qtl.sh
QTLtools cis --vcf geno.vcf.gz --bed expression.bed.gz --permute 1000    tabix -p vcf imputed.chr10.dose.vcf.gz
samtools view -b -q 30 -F 1804 aln.bam chr10:112990000-113010000    bcftools norm -m -any -f GRCh38.fa | bcftools view -q 0.01:minor
cellranger-atac count --reference GRCh38 --fastqs runs/ --id muscle_01    liftOver credible_sets.hg19.bed hg19ToHg38.over.chain.gz out.bed unmapped.bed
STAR --runThreadN 16 --genomeDir GRCh38 --readFilesIn R1.fq.gz R2.fq.gz    ldsc.py --h2-cts muscle.sumstats.gz --ref-ld-chr baseline. --w-ld-chr weights.
nextflow run main.nf -profile slurm,singularity -resume    macs2 bdgcmp -t treat.bdg -c control.bdg -m FE -o fold_enrichment.bdg
snakemake --profile slurm --use-singularity --jobs 400    gzip -dc gwas.sumstats.gz | sort -k1,1 -k2,2n | bgzip > sorted.gz
sbatch --array=1-402%40 --mem=32G --time=24:00:00 map_qtl.sh    bedtools intersect -a snATAC_peaks.bed -b credible_sets.bed -wa -wb -u
tabix -p vcf imputed.chr10.dose.vcf.gz    macs2 callpeak -t muscle.nuclei.bam -f BAMPE -g hs --nomodel --shift -100
bcftools norm -m -any -f GRCh38.fa | bcftools view -q 0.01:minor    plink2 --pfile cohort --glm hide-covar --covar covariates.tsv --out eqtl
liftOver credible_sets.hg19.bed hg19ToHg38.over.chain.gz out.bed unmapped.bed    Rscript -e 'susieR::susie_rss(z, R, n = 402, L = 10)'
ldsc.py --h2-cts muscle.sumstats.gz --ref-ld-chr baseline. --w-ld-chr weights.    Rscript -e 'coloc::coloc.abf(gwas, qtl, p12 = 1e-5)'
macs2 bdgcmp -t treat.bdg -c control.bdg -m FE -o fold_enrichment.bdg    QTLtools cis --vcf geno.vcf.gz --bed expression.bed.gz --permute 1000
gzip -dc gwas.sumstats.gz | sort -k1,1 -k2,2n | bgzip > sorted.gz    samtools view -b -q 30 -F 1804 aln.bam chr10:112990000-113010000
Arushi Varshney

I work out how genetic variation changes gene regulation at the University of Michigan. Single-nucleus multi-omics across hundreds of human samples, mapped down to the cell types, regulatory elements and genes behind metabolic disease. Before that, a PhD in human genetics here and bioengineering at EPFL.
A GWAS signal names a stretch of the genome, not a mechanism. The work is finding which variant, in which cell type, acting on which gene.
Fig. 1 Two signals, one locus. When a trait association and a molecular QTL rise and fall together, the variant they share is the one worth chasing into a cell type.