GTEx_edQTL

This depository contains code used to produce cis RNA editing QTL (edQTL) mapping results using the v8 GTEx data.

The edQTL mapping data is deposited at https://gtexportal.org.

Quantification of RNA editing levels

To quantify RNA editing levels in the v8 GTEx RNA-seq data, use Docker to run parse_pileup_query.pl

Combine individual editing data into a matrix

Run the sharedsamples_sites_matrix_FastQTL_v8.pl script to convert individual editing data files into a matrix. Change minsamps in the script to set the minimum number of samples per site (default is 60 samples). Change mincov to set the minimum reads coverage per site (default is 20 reads).

perl sharedsamples_sites_matrix_FastQTL_v8.pl > ${tissue}.edMat.20cov.60samps.txt

Convert editing level matrices to format recognized by FastQTL

Use the prepare_phenotype_table_for_QTLtools.V8.py script to convert. This step will also perform quantile normalization of editing level across samples. tabix is also required.

python /labs/lilab/qin/GTEx/scripts/prepare_phenotype_table_for_QTLtools.V8.py -p 15 ${tissue}.edMat.20cov.60samps.noXYM.txt
sh ${tissue}.edMat.20cov.60samps.noXYM.txt_prepare.sh
zcat ${tissue}.edMat.20cov.60samps.noXYM.txt.qqnorm_chr*gz | sed 's/nan/NA/g' | sort -k 1,1 -k2n,2 | awk '{print "chr"$1,$2,$3,"chr"$1"_"$3,$0}' OFS="\t" | cut -f 1,2,3,4,9- | sed 's/^chr#/#/g' | bgzip -c > ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.bed.gz
tabix -f -p bed ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.bed.gz

Obtain PEER factors from the combined editing level matrices

Obtain PEER factors of the editing levels using run_PEER.R. This step will output ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.PEER_covariates.txt file containing the top 60 PEER factors.

Rscript /labs/lilab/qin/software/gtex-pipeline/qtl/src/run_PEER.R \
    ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.bed.gz \
    ${tissue}.edMat.20cov.60samps.noXYM.qqnorm \
    60

Combine genotype PCs with phenotype PEER factors

The genotype PCs are obtained from the genotyping VCF file (generated by the GTEx consortium). Also include sex and age information as additional covariates.

python combine_covariates.py \
    ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.PEER_covariates.txt \
    ${tissue}.edMat.20cov.60samps.noXYM.qqnorm \
    -o ${tissue} \
    --add_covariates GTEx_Analysis_2017-06-05_v8_Annotations_SubjectPhenotypesDS.SEX_AGE.txt \
    --genotype_pcs GTEx_Analysis_2017-06-05_v8_WholeGenomeSeq_838Indiv_Analysis_Freeze_20genotPCs.txt

Running FastQTL: nominal step

Running the cis QTL mapping nominal step with cis-window = 100kb. Genotyping information in vcf format is generated by the GTEx consortium.

python3 run_FastQTL_threaded.py \
    GTEx_Analysis_2017-06-05_v8_WholeGenomeSeq_838Indiv_Analysis_Freeze_MAF005_GTonly.vcf.gz \
    ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.bed.gz \
    ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.nominal \
    --covariates ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.combined_covariates.txt \
    --maf_threshold 0.05 \
    --ma_sample_threshold 4 \
    --threads 4 \
    --window 1e5

Adaptive permutaion run

python3 run_FastQTL_threaded.py \
    GTEx_Analysis_2017-06-05_v8_WholeGenomeSeq_838Indiv_Analysis_Freeze_MAF005_GTonly.vcf.gz \
    ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.bed.gz \
    ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.perm \
    --covariates ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.combined_covariates.txt \
    --permute 1000 10000 \
    --ma_sample_threshold 4 \
    --maf_threshold 0.05 \
    --chunks 100 \
    --threads 8 \
    --window 1e5

Annotating the results and generating edSites information

Need the GTEx_Analysis_2017-06-05_v8_WholeGenomeSeq_838Indiv_Analysis_Freeze.lookup_table.txt.gz to provide annotation of the variants (generated by the GTEx consortium)

snp_lookup=GTEx_Analysis_2017-06-05_v8_WholeGenomeSeq_838Indiv_Analysis_Freeze.lookup_table.txt.gz
gtf_file=All.AG.stranded.annovar.Hg38_multianno.AnnoAlu.AnnoRep.NR.gtf
python3 annotate_outputs.py \
    --snp_lookup=$snp_lookup \
    --nominal_results ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.nominal.txt.gz \
    -o ${tissue} \
    ${tissue}.edMat.20cov.60samps.noXYM.qqnorm.perm.txt.gz \
    0.05 \
    $gtf_file

Please contact Qin Li qinl@stanford.edu if there is any question.