Figure 2. Assessment of the association of nucDNA and mtDNA loci contributing to the mitochondrial proteome with age-related traits.
(A) Scheme outlining the aspects of mitochondrial function assessed in this study. nucDNA loci contributing to the mitochondrial proteome are shown in teal, while mtDNA loci are shown in pink. (B) S-LDSC enrichment p-values on top of the baseline model in UKB. Inset labels represent gene-set size; dotted line represents BH FDR 0.1 threshold. (C) Visualization of mtDNA variants and associations with age-related diseases. The outer-most track represents the genetic architecture of the circular mtDNA. The heatmap track represents the log-scaled number of individuals with an alternate genotype at each site. The inner track represents mitochondrial genome-wide association p-values, with radial angle corresponding to position on the mtDNA and magnitude representing –log10 p-value. Dotted line represents Bonferroni cutoff for all tested trait-variant pairs. (D) Replication of S-LDSC enrichment results in meta-analyses. Dotted line represents BH FDR 0.1 threshold. * represents traits for which sufficiently well-powered cohorts from both UKB and meta-analyses were available. The trait color legend to the right of panel (C) applies to panels (B) and (C), representing UKB traits. S-LDSC enrichment p-values plotted in (B) and (D) are available in Source data 1; mtDNA-GWAS summary statistics are available in Source data 2.
Figure 2—figure supplement 1. Enrichment tests in mitochondria-localizing genes in the GWAS Catalog.
(A) The proportion of annotated lead SNP-associated genes for manually curated phenotypes from the GWAS Catalog that correspond to age-related traits of interest (colors) that overlap the set of mitochondria-localizing genes. Black line represents the proportion of all genes in the genome that are mitochondria-localizing, with bars above the black line representing nominal enrichment and below representing nominal depletion. * represents traits showing significant depletion or enrichment via two-sided Fisher’s exact test at BH FDR q-value < 0.1. (B) Empirical null distribution of the test statistic of the number of traits with nominal enrichment computed via 1000 randomly sampled subsets of the set of all protein-coding genes. Dotted line represents the true observed value for mitochondria-localizing genes. (C) Rank plot of the proportion of trait-associated genes overlapping with mitochondria-localizing genes (the same values as represented on the x-axis of panel (A) across all GWAS catalog phenotypes with at least 30 trait-associated genes). Dark points indicate traits with a significant depletion or enrichment via two-sided Fisher’s exact test at BH FDR q-value < 0.1. Dotted line represents the proportion of all genes in the genome that are mitochondria-localizing, with traits above the black line representing nominal enrichment and below representing nominal depletion. No traits show significant enrichment.
Figure 2—figure supplement 2. Analysis of mitochondria-localizing gene enrichments using alternative methods and cohorts.
(A) Coefficient point estimates and corresponding 95% CI for mitochondria-localizing gene-sets in UKB with p-values shown in Figure 2B. (B) Coefficient point estimates and 95% CI for mitochondria-localizing gene-sets tested for replication in meta-analyses with p-values shown in Figure 2D. MAGMA enrichments for mitochondria-localizing genes with 5 kb up and 1.5 kb down gene window in (C) UKB and (D) meta-analyses. (E) MAGMA enrichments for mitochondria-localizing genes in UKB with a 100 kb symmetric gene window. Dotted line represents BH FDR 0.1 threshold. * represents traits for which sufficiently well-powered cohorts from both UKB and meta-analyses were available. Inset numbers represent gene-set sizes. Enrichment p-values and effect sizes are available in Source data 1.
Figure 2—figure supplement 3. S-LDSC enrichments for tissues across age-related traits.
Enrichment P-values in (A) UK Biobank and (B) meta-analyses using S-LDSC on top of the baseline model. Gene-sets of the top 10% tissue-specific genes for each tested tissue were obtained using GTEx v7 as described previously (Finucane et al., 2018), with tissues for analysis selected manually based on the set of tested traits. * represents traits for which sufficiently well-powered cohorts from both UKB and meta-analyses were available. Dashed line represents –log10 p-value cutoff for Benjamini-Hochberg FDR < 10%. Enrichment p-values are available in Source data 1.
Figure 2—figure supplement 4. S-LDSC enrichment coefficients from tissue analysis.
(A) Coefficient point estimates and corresponding 95% CI for tissue-expressed gene-sets in UKB with p-values shown in Figure 2—figure supplement 3. (B). Coefficient point estimates and 95% CI for tissue-expressed gene-sets tested in meta-analyses with p-values shown in Figure 2—figure supplement 3. Inset numbers represent gene-set sizes. * represents traits for which sufficiently well-powered cohorts from both UKB and meta-analyses were available. All analyses were performed atop the baseline model. Enrichment coefficients and standard errors are available in Source data 1.
Figure 2—figure supplement 5. MAGMA enrichments for tissues in UKB.
Enrichment P-values for tissues in UKB with a (A) 5 kb up and 1.5 kb down window and (B) 100 kb symmetric window. These gene-sets are analogous to those shown tested with S-LDSC in Figure 2—figure supplement 3 and Figure 2—figure supplement 4. Inset numbers represent gene-set sizes. Dashed line represents –log10 P-value cutoff for Benjamini-Hochberg FDR < 10%. * represents traits for which sufficiently well-powered cohorts from both UKB and meta-analyses were available. Enrichment p-values and effect sizes are available in Source data 1.
Figure 2—figure supplement 6. MAGMA enrichments for tissues in meta-analyses.
Enrichment p-values for tissues in meta-analyses with a (A) 5 kb up and 1.5 kb down window and (B) 100 kb symmetric window. These gene-sets are analogous to those shown tested with S-LDSC in Figure 2—figure supplement 3 and Figure 2—figure supplement 4 and with MAGMA in Figure 2—figure supplement 5. Dashed line represents –log10 p-value cutoff for Benjamini-Hochberg FDR < 10%. No p-values crossed the FDR threshold for the 100 kb window, and as such the BH p-value cutoff cannot be displayed. * represents traits for which sufficiently well-powered cohorts from both UKB and meta-analyses were available. Enrichment p-values and effect sizes are available in Source data 1.
Figure 2—figure supplement 7. Enrichment power of S-LDSC and MAGMA across effect sizes and gene-set sizes.
Using biologically plausible trait-tissue associations originally computed at gene-set sizes of ~2500 genes (Atrial Fibrillation – Atrial appendage; Diverticular Disease – Sigmoid colon; Glucose, T2D – Pancreas; HDL, LDL, Cholesterol, TG – Liver, HDL – Visceral adipose; Myocardial Infarction – Coronary artery), power was computed as a probability of rejection at P ≤ 0.05 across 1000 randomly sampled subsets (of size 1520, 1105, 800, and 350 genes) of the original tissue expression gene-set. (A) Relationship between S-LDSC coefficient magnitude and power for 1105 genes. (B) Relationship between MAGMA effect size magnitude and power for 1105 genes. (C) Relationship between S-LDSC coefficient magnitude and power for 350 genes. (D) Relationship between MAGMA effect size magnitude and power for 350 genes. Shape represents window size, color represents disease, and small points represent trait-tissue pairs that were not detected as enriched at the full gene-set. (E) Power as a function of gene-set size for MAGMA and S-LDSC for a very high effect size trait-tissue pair: LDL levels and liver. (F) Power as a function of gene-set size for MAGMA and S-LDSC for a low effect size trait-tissue pair: Atrial fibrillation and atrial appendage. UKB vs. meta-analysis power comparisons for (G) 1523 genes, (H) 1105 genes, (I) 800 genes, and (J) 350 genes.
Figure 2—figure supplement 8. Assessment of cis-eQTLs among mitochondria-localizing genes.
(A) Percentage of genes localizing to each organelle that have an observed cis-eQTL in at least one GTEx tissue. (B) Coefficient effect size estimates for the regression of the number of distinct tissues with detected cis-eQTL for a given gene onto an indicator for organelle membership (Materials and methods). (C) Replication of (B) with alternate selection of tissues within each tissue class for robustness (Appendix 1). Error bars represent 95% CI.
Figure 2—figure supplement 9. Landscape of mtDNA variants and their associations with age-related disease in UKB.
(A) Within-mtDNA pair-wise correlation coefficients across all 213 tested variants with expected case count of alternative genotype individuals > 20. Axis labels represent mtDNA coordinates, color represents . (B) Quantile-quantile plot of p-values for all tested variant-trait pairs (21 traits x ~213 variants = ~4473 tests) using linear regression for all traits while controlling for baseline characteristics (first 20 PCs of nuclear genotype matrix, age, sex, age*sex, age2*sex) as well as array type. (C) Quantile-quantile plot of all tested variant-trait pairs for all traits using linear regression, controlling only for baseline characteristics. (D) Quantile-quantile plot of all tested variant-trait pairs using linear regression for continuous traits and logistic regression for binary traits, controlling only for baseline characteristics. Red line represents expected null p-values following the uniform distribution and shaded ribbon represents 95% CI. (E) Visualization of mtDNA variants and associations with serum creatinine and aspartate aminotransferase levels. The outer-most track represents the genetic architecture of the circular mtDNA. The heatmap track represents the number of individuals with alternate genotype on log scale. The inner track represents mitochondrial genome-wide association p-values, with radial angle corresponding to position on the mtDNA and magnitude representing –log10 p-value. Dotted line represents Bonferroni cutoff for all tested trait-variant pairs, using the same threshold as used in Figure 2C. mtDNA-GWAS summary statistics are available in Source data 2.