Supplementary data files associated with this manuscript.
(a) 2156 protein-coding genes that are significant by S-MultiXcan transcriptome-wide association study (TWAS) analysis (Bonferroni p value ≤0.05). Columns are Ensembl ID, gene name, the nominal S-MultiXcan p value, the number of tissues available to S-MultiXcan, the number of independent components of variation among the tissues, and the Bonferroni-adjusted p value. (b) 1182 protein-coding genes that are significant in the fastEnloc colocalization analysis (regional colocalization probability [RCP] ≥0.1). Columns are Ensembl ID, signal cluster name from the expression quantitative trait loci (eQTL) analysis, the gene name, the RCP, and the colocalizing Genotype-Tissue Expression (GTEx) tissue. (c) The 512 protein-coding genes that are significant by both TWAS and colocalization (Bonferroni p value ≤0.05 and RCP ≥0.1). Columns are Ensembl ID, signal cluster name from the eQTL analysis, gene name, the RCP, the colocalizing GTEx tissue, the nominal S-MultiXcan p value, and the Bonferroni-adjusted p value. (d) Number of the 512 significantly colocalizing genes per GTEx tissue. Columns are the GTEx tissue and the number of unique colocalizing eQTL in the relevant tissue. (e) The ‘known bone gene’ list. Columns are gene name and Ensembl ID. (f) The 66 genes that are significant by both TWAS and colocalization (Bonferroni p value ≤0.05 and RCP ≥0.1), and are also members of the ‘known bone gene’ list. Columns are gene name and Ensembl ID. (g) Gene Ontology (GO) enrichments for the 512 protein-coding genes that are significant by both TWAS and colocalization (Bonferroni p value ≤0.05 and RCP ≥0.1). Only GO enrichments with a p value ≤0.05 are included. Columns are the GO IDs, GO terms, p values, the GO subontology (BP – Biological Process, CC – Cellular Component, MF – Molecular Function), and the Ensembl IDs for the genes that are members of the GO ontology. p values were calculated using a one-sided Fisher’s exact test, and were not adjusted for multiple comparisons. (h) The 863 genes closest to estimated bone mineral density (eBMD) genome-wide association study (GWAS) associations. These data were obtained from Morris et al., 2019. (i) 137 novel putatively causal BMD genes, after increasing RCP ≥0.5, and removing genes that were members of the ‘known bone gene’ list and genes with a nominal (p ≤ 0.05) alteration in BMD as determined by the International Mouse Phenotype Consortium (IMPC). Columns are Gene name, Ensembl ID, colocalization RCP, the tissue with the highest RCP, the S-Multixcan Bonferroni-adjusted p value, and the number of tissues that were significantly colocalizing. Note that TLN2 appears twice due to having the same RCP in two tissues. (j) Results of the differential expression analysis performed on RNA isolated from vertebrae. The analysis results show the difference in expression between mutant and wild-type mice. The columns are the Ensembl IDs, gene names, baseMean (the average of normalized count values, taken over all samples), log2 fold change, standard error of the log2 fold change estimate, the Wald statistic, the test p value, and the BH-adjusted p value.