Expression QTLs for NYGC ALS Consortium Paper

The files below contain nominal and permuted quantitative trait loci (QTL) associations between common genetic variants derived from whole genome sequencing and gene expression phenotypes generated from RNA-seq of post-mortem tissue sections. All QTLs were mapped with TensorQTL.

Top association files are gzip-compressed tab-separated variable files - *cis_qtl.txt.gz

Nominal association files are stored as Parquet files to save space. These can be converted to text files using the following code snippet:

pip install pandas pyarrow
conda install -c bioconda htslib  # provides bgzip

python3 -c "
import pandas as pd
df = pd.read_parquet('your_file.parquet')
df.to_csv('your_file.tsv.gz', sep='\t', index=False)
" | bgzip > your_file.tsv.gz

File Descriptions

Table columns are formatted as follows:

Nominal QTL results include all SNP-gene pairs tested using either a 1Mb window from each side of the transcription start site (TSS) of the gene. 

• phenotype_id - ensembl ID of the gene tested (GENCODE v30)
• variant_id - SNP tested for association (rsid or chr:position:ref:alt)
• tss_distance - distance of the SNP to the gene transcription start site (TSS)
• maf - minor allele frequency in cohort
• ma_samples - number of samples carrying the minor allele
• ma_count - total number of minor alleles across individuals
• pval_nominal - nominal P-value from linear regression
• slope - slope of the linear regression
• slope_se - standard error of the slope

Top association results include only the top SNP-gene association for each gene. Table columns are formatted as follows:

• phenotype_id - ensembl ID of the gene tested (GENCODE v30
• num_var - total number of variants tested in cis
• beta_shape1 - first parameter value of the fitted beta distribution
• beta_shape2 - second parameter value of the fitted beta distribution
• true_df - effective degrees of freedom the beta distribution approximation
• pval_true_df - empirical P-value for the beta distribution approximation
• variant_id - ID of the top variant (rsid or chr:position:ref:alt)
• tss_distance - distance of the SNP to the gene transcription start site (TSS)
• ma_samples - number of samples carrying the minor allele
• ma_count - total number of minor alleles across individuals
• maf -minor allele frequency in MiGA cohort
• ref_factor - flag indicating if the alternative allele is the minor allele in the cohort (1 if AF <= 0.5, -1 if not)
• pval_nominal - nominal P-value from linear regression
• slope - slope of the linear regression
• slope_se - standard error of the slope
• pval_perm - first permutation P-value directly obtained from the permutations with the direct method
• pval_beta - second permutation P-value obtained via beta approximation. This is the one to use for downstream analysis
• qval - Storey q-value derived from pval_beta (FDR adjusted)
• pval_nominal_threshold - nominal P-value threshold for calling a variant-gene pair significant for the gene

Allele Information for each variant NYGC_all_common_variants_alleles.tsv.gz

• CHROM - chromosome position of the variant
• POS - position of the variant in the chromosome
• REF - reference allele (GRCh38)
• ALT - alternative allele (this is the effect allele in the eQTL analysis)
• ID - variant id (rsid or chr:position:ref:alt)

Covariates files <tisssue>_expression_eQTL_covariates.txt.gz

Wide format for tensorQTL. Each column is a donor. The following covariates are included:

• PC1-5 - principal components of the genotypes
• wgs_prep - library preparation method for WGS
• age_at_death - age of death
• rna_prep - library preparation method for RNA-seq
• biological_sex - the biological sex of each donor
• InferredCov1-15 - PEER factors generated on the expression matrix. Larger tissues have 15, smaller have 10.

 Sample keys <tissue>_sample_key.txt

• sample_id - the RNA-seq sample
• participant_id - the WGS donor