TWAS EUR weights

The EUR gene expression models of 5490 genes with significant, positive h2 estimated by GCTA. Note not extracting 1KG EUR hapmap3 variants.

TWAS AFR weights

The AFR gene expression models of 3244 genes with significant, positive h2 estimated by GCTA. Note not extracting 1KG EUR hapmap3 variants.

zenodo v3

zenodo_v3

updated COLOC and eCAVIAR code

zenodo v2

zenodo_v2

Update README.md

Gene counts

• gene.noVersion.rawCounts.tsv.zip: raw gene counts, quantified by Salmon, compiled by tximport
• genes.combatseq.tsv.zip: raw gene counts, restricted to gene and sample sets used in eQTL mapping (31947 genes, 642 samples), combat-seq to correct for batch. WGCNA network analysis starts from this file, and then TMM normalize to generate a pseudo-count matrix of normalized expression

Genotypic variants

test.bim.zip: Genotypic variants for multi-ancestry xQTL mapping. A1 is the effect allele.
filtered.hg19.sorted.removeRel.bim.zip: Genotypic variants for EUR xQTL mapping. A1 is the effect allele

zenodo

Make zenodo archive

Supplementary Data for medRxiv

Supplementary Data
Data S1 (Mixed_ciseqtl_90hcp_perm_purity_filtered.txt.gz): SuSiE fine-mapping results for multi-ancestry cis-eQTL
Data S2 (Eur_ciseqtl_50hcp_perm_purity_filtered.txt.gz): SuSiE fine-mapping results for EUR cis-eQTL
Data S3 (Amr_ciseqtl_15hcp_perm_purity_filtered.txt.gz): SuSiE fine-mapping results for AMR cis-eQTL
Data S4 (Afr_ciseqtl_25hcp_perm_purity_filtered.txt.gz): SuSiE fine-mapping results for AFR cis-eQTL
Data S5 (EUR_tri1_gene_cish2.csv): heritability estimates for genes in EUR Tri1 samples
Data S6: EUR_tri2_gene_cish2.csv): heritability estimates for genes in EUR Tri2 samples
Data S7: (EUR_tri1_splicing_cish2.csv): heritability estimates for local splicing in EUR Tri1 samples
Data S8: (EUR_tri2_splicing_cish2.csv): heritability estimates for local splicing in EUR Tri2 samples
Data S9 (modeqtl_top.RData): module interaction eQTL permutation eGenes and their primary ieQTLs

Supplementary Tables for medRxiv

Table S1. Metadata and xQTL summary statistics
● README
● ST1-1-metadata: Metadata of 654 distinct, high-quality samples
● ST1-2-eGene: Sum stats of 10094 permutation eGenes and their primary cis-eQTL
● ST1-3-isoGene: Sum stats of 11845 group permutation isoGenes and their primary cis-isoQTL
● ST1-4-sGene: Sum stats of 7490 group permutation sGenes and their primary cis-sQTL
● ST1-5-torus-funcEnrich: Torus functional enrichment of cis-e/iso/sQTL
● ST1-6-conditional: Sum stats of conditional cis-eQTL mapping, for detailed descriptions of columns, see https://qtltools.github.io/qtltools/
● ST1-7-fetalOnlyeGene-CellTypeEnrich: Cell type enrichment for 2488 fetal only eGenes. "apprach" column denotes whether the cell types are defined as broad, cluster, or sub-cluster
● ST1-8-inv-eQTL: Sum stats of significant inversion eQTL

Table S2. Population specific xQTL and fine-mapping
● README
● ST2-1-EUR-eGenes: Sum stats of 5296 EUR permutation eGenes and their primary cis-eQTL
● ST2-2-AMR-eGenes: Sum stats of 3242 EUR permutation eGenes and their primary cis-eQTL
● ST2-3-AFR-eGenes: Sum stats of 1876 EUR permutation eGenes and their primary cis-eQTL
● ST2-4-EUR-isoGenes : Sum stats of 11672/6885 isoforms/isoGenes with permutation cis-isoQTL in EUR
● ST2-5-AMR-isoGenes: Sum stats of 5747/3989 isoforms/isoGenes with permutation cis-isoQTL in AMR
● ST2-6-AFR-isoGenes: Sum stats of 3719/2682 isoforms/isoGenes with permutation cis-isoQTL in AFR
● ST2-7-EUR-sGenes: Sum stats of 16277/4963 introns/sGenes with permutation cis-sQTL in EUR
● ST2-8-AMR-sGenes: Sum stats of 6324/2772 introns/sGenes with permutation cis-sQTL in AMR
● ST2-9-AFR-sGenes: Sum stats of 4337/2069 introns/sGenes with permutation cis-sQTL in AFR

Table S3. Trimester xQTL summary statistics
● README
● ST3-1-tri1-eGene: Sum stats of 4211 permutation eGenes in Tri1
● ST3-2-tri2-eGene: Sum stats of 2220 permutation eGenes in Tri2
● ST3-3-tri1-isoGene: Sum stats of 10881/6921 isoforms/isoGenes with permutation cis-isoQTL in Tri1
● ST3-4-tri2-isoGene: Sum stats of 4179/3047 isoforms/isoGenes with permutation cis-isoQTL in Tri2
● ST3-5-tri1-sGene: Sum stats of 14193/5312 introns/sGenes with permutation cis-sQTL in Tri1
● ST3-6-tri2-sGene: Sum stats of 5348/2318 introns/sGenes with permutation cis-sQTL in Tri2
● ST3-7-triOnly-biotype: Counts of gene types of tri1 and tri2-only e/sGenes
● ST3-8-triOnly-cellTypeEnrich: Cell type enrichment of tri1 and tri2-only e/sGenes

Table S4. S-LDSC and MESC results
● README
● ST4-1-sLDSC: S-LDSC of background and various brain QTL annotations (e/iso/sQTL, GTEx brain cortex eQTL, tri e/iso/sQTL)
● ST4-2-sLDSC-e/iso/sQTL-joint: S-LDSC jointly running background annotations, and e+iso+sQTL, top enrichment 15 annotations
● ST4-3-sLDSC-e/isoQTL-joint: S-LDSC jointly running background annotations, and e+isoQTL, top enrichment 15 annotations
● ST4-4-MESC: h2med results for all expressed genes/isoforms/introns in the full dataset, and trimester specific datasets

Table S5. isoTWAS and colocalization
● README
● ST5-1a-isoTWAS-preFOCUS: isoTWAS statistics of isoforms that are permutation significant
● ST5-1b-isoTWAS-finemapped: isoTWAS statistics of isoforms that are in 90% FOCUS credible sets
● ST5-2 to ST5-29: eCAVIAR significant colocalization results for various neuropsychiatric GWAS and fetal brain QTLs
● ST5-30: MRLOCUS summary statistics

Table S6. Network analyses
● README
● ST6-1-Module-Genes: Membership of genes and transcripts to each of the modules
● ST6-2-All-Enrichment: Compiled module enrichment data from ST6-3, ST6-4, and ST6-5
● ST6-3-Cell-Enrichment: Cell type enrichment of modules using Polioudakis fetal brain marker genes
● ST6-4-Common-Enrichment: Common variation enrichment of modules using sLDSC and MAGMA
● ST6-5-Rare-Enrichment: Rare variation enrichment of modules using logistic regression
● ST6-6-Module-Overlap: Overlap of fetal modules with other fetal modules in the dataset or modules from Walker et al. (walker_), Werling et al. (werling_), and Li et al. (li_)

Table S7. CellWalker and module interacting eQTL summary statistics
● README
● ST7-1-CellWalker-single-label: CellWalker mapping of bulk eQTL SuSiE results to single cell type labels. Corresponds to Fig 7A
● ST7-2-CellWalker-ML-strict: CellWalker mapping of bulk eQTL SuSiE results to multi-level ("ML") cell type labels. Filtered for number of cell types. See methods for details

TWAS Weights

Gene-level TWAS weights:

1. PredictDB data format for PrediXcan family of methods
2. ZIP of individual gene .wgt files for TWAS / Fusion Methods