Developing brain mega xQTL

Data sharing

• Final supplemental tables available with the paper.
• GitHub releases: TWAS/PrediXcan weights, SuSiE fine-mapping results.
• Galaxy: https://usegalaxy.org/u/cindywen/h/devbrainsumstats. Cross-ancestry e/iso/sQTL nominal and permutation summary statistics, effect alleles, MAFs.
• Synapse: https://doi.org/10.7303/syn50897018.5. QTL summary statistics. Requires sign-in.
• Gandal lab internal: Google doc with paths to fils on Hoffman2.

1: Munging data

2: Genotype

2-1: Pre-imputation

• Run plinkQC on data as a sanity check
• First apply PLINK filters, then split by chromosome and sort
  • Walker data is already filtered; split by chromosome and impute
  • For all the other datasets, we applied the same filters that the Walker data used --hwe 1e-6 --maf 0.01 --mind 0.10 --geno 0.05
  • Note: for HDBR, we used --mind 0.3; for LIBD, we fixed strand flips by running an extra step of conform-gt, which automatically splits the data by chromosome

2-2: Post-imputation

• Scripts in prelim/: inputs are imputed genotype files downloaded from Michigan Imputation Server; concatenate by chromosomes, index, filter by R2, and take the intersection of high impute quality variants across datasets
  • Note: except for Walker data, we applied R2>.3 filter during imputation; so here we only applied R2>.3 on Walker imputed data and intersected with the other datasets
• ancestry.ipynb: infer data ancestry, make plots
• IBD.ipynb: relatedness check
• Snakefile

3: RNA-seq

• Pre-alignment QC FastQC v0.11.9
• Alignment STAR-2.7.3a, index with GENCODE v29lift37 genome and annotation
  • Note: there is a new run of STAR for sQTL
• Alignment QC PicardTools 2.21.7
• Compile FastQC and PicardTools metrics MultiQC v1.9.dev0

# In picard/
# -d -dd 1: to keep identical sample ID from different folders
python3 -m multiqc -d -dd 1 Walker/ Obrien Werling_final/ hdbr libd -o all_multiqc

• Quantification Salmon v1.1.0, GENCODE v33lift37 decoys-aware index
• Compile and import quantifications Tximport 1.14.0

txi <- tximport(files, type="salmon", tx2gene=tx2gene, dropInfReps=TRUE, countsFromAbundance="lengthScaledTPM")
write.table(txi$counts,file="gene.noVersion.scaled.counts.tsv",quote=FALSE, sep='\t')
write.table(txi$abundance,file="gene.noVersion.TPM.tsv",quote=FALSE, sep='\t')

txi.tx <- tximport(files, type="salmon", txOut=TRUE, dropInfReps=TRUE, countsFromAbundance="lengthScaledTPM")
write.table(txi.tx$counts,file="tx.counts.scaled.tsv",quote=FALSE, sep='\t')
write.table(txi.tx$abundance,file="tx.TPM.tsv",quote=FALSE, sep='\t')

• Sample swap check:
  • VerifyBamID (slow. Use --smID to add subject ID to BAM sequence file)
  • check.ipynb: called SNP from BAM, merged with imputed genotype (Mike)

4: xQTL

4-1: cis-eQTL

• ancestry.ipynb
• combat-seq.ipynb
• decon.ipynb: cell type specific and interacting analysis
• eqtl_analysis.ipynb: identify optimal #HCP in covariates, gene expression PCA, dTSS, etc.
• fetal_adult.ipynb
• func_enrich.ipynb: functional enrichment analysis of QTL
• metadata.ipynb: plot data age, sex, infer NA sex, etc.
• module_eigengene.ipynb
• paintor.ipynb: PAINTOR multi-ethnic fine-mapping
• pLI.ipynb
• sex_specific.ipynb
• susie.ipynb: susie finemapping results
• tri_egene_biotype.ipynb
• tri_h2_supp.ipynb
• tri_specific.ipynb
• walker_fetal.ipynb
• Snakefile
• decon.smk
• paintor.smk

4-2: cis-isoQTL

• isoqtl_analysis.ipynb
• prep.ipynb: sex and trimester specific QTL
• Snakefile: follows a similar pipeline as cis-eQTL, except that run grouped permutation as GTEx did

4-3: cis-sQTL

• sqtl_analysis.ipynb
• e_iso_s.ipynb
• qvalue_pi0.ipynb
• check.ipynb: check chunk size
• Snakefile

4-4: trans

• gbat.ipynb
• Snakefile

4-5: APEX

• apex_analysis.ipynb
• Snakefile

5: Integrative analysis

See coloc_ecaviar_May_2024/ for colocalization analysis

5-1: sLDSC

• ldsc_analysis.ipynb
• Snakefile
• pec.smk

5-2: TWAS-FUSION

• TWAS.ipynb
• LDREF.ipynb
• run_focus.sh
• Snakefile

5-3: MESC

• MESC.ipynb
• Snakefile
• test.smk

5-4: Colocalization (eCAVIAR)

• eCAVIAR.ipynb
• GRIN2A.ipynb
• SP4_gviz.ipynb
• sqtlviztools.ipynb
• Visualizing_Loci_working.ipynb
• celltype.smk
• eqtl.smk
• isoqtl.smk
• mod_ieqtl.smk
• sex_tri.smk
• sqtl.smk
• sashimi plot related code

6: Further Analyses

6-1: eGene/sGene Enrichment

• fetal_only_egenes.ipynb: biotype and cell type analysis for fetal-specific eGenes
• trimester_egenes_sgenes.ipynb: biotype and cell type analysis for trimester-specific e/sGenes

6-2: WGCNA

• compare_module_enrichment.ipynb: compare enrichment across networks and across correlated cell types
• dashboard_generator.ipynb: generate dashboards using ST6.xlsx
• dashboards: folder containing dashboards for each module