Mitochondrial genetic variation across tissues of the human body

Figures

• 

  Figure 1.. Overview of the identified mtDNA variants. Distribution of mtDNA variants by VAF (%) in the innermost track, gene in the middle track, and position in the outermost track (A). Number of total and unique mtDNA variants identified in the entire GTEx dataset stratified by variant class across tissues within donors (B). Percent of donors with a given number of variants (1, 2–10, 11–30, and >30) stratified by variant class (C). Proportions of fully shared, partially shared, and tissue-specific variants in OXPHOS complex (I, III, IV, V) and rRNA genes (D). Transition to transversion ratios for mtDNA variants (E). dN/dS for mtDNA variants (F). Mean MLC scores for mtDNA variants (G). Proportions of synonymous, nonsynonymous, nonstop, and nonsense variants among fully shared, partially shared, and tissue-specific mtDNA variants (H). Distribution of VAFs for the identified mtDNA variants (I).

• 

  Figure 2.. Frequency and functional consequences of tissue-specific variants vary by tissue. Mean number of tissue-specific variants per sample by tissue. Error bars show ± 1 standard error from the mean (A). Heatmap of the fold-difference (log2-transformed) in mean number of tissue-specific variants between different tissues (B). dN/dS for each tissue stratified by variant class (C). Black dashed line ind icates a y-axis break. Mean MLC scores for the variants for each tissue type (D). Red dashed line at MLC score = 0.5 indicates the cutoff for variants considered to be deleterious (MLC score > 0.5). Heatmap of the fold-difference (log2-transformed) in mean MLC score between tissue types (E). Distributions of the five most common recurrent tissue-specific variants by tissue type (F). A single block indicates one donor. *Significant Dunn post-hoc test comparisons with Benjamini–Hochberg adjusted P <.05.

• 

  Figure 3.. Variant type proportions associated with mtDNA polymerase gamma variant strand bias. Proportions of transition and transversion types stratified by strand and variant class (A) and stratified by heteroplasmic versus homoplasmic variants (B). Tissue-specific variant proportions by type of transitions and transversions stratified by strand and tissue for the 10 tissues with the highest numbers of tissue-specific variants (C).

• 

  Figure 4.. Variants with confirmed pathogenicity are typically tissue-specific variants. Frequency and VAF of mtDNA variants in the GTEx dataset that are confirmed pathogenic variants in MITOMAP [35]. The bar plot indicates the number of samples with a given variant, colored by tissue type. Variants are divided by those that are tissue-specific and those that exist in multiple tissues for a given donor (multi-tissue).

• 

  Figure 5.. Frequency of tissue-specific variants varies within organs and tissues. The relative proportions of sampling sites within a tissue/organ with at least one fully shared, partially shared, or tissue-specific variant and the dN/dS values for each sampling site for esophagus (A), skin (B), and heart (C). Percentage of brain tissue samples by sampling site that contain at least one tissue-specific variant (D).

• 

  Figure 6.. Developmental origins of partially shared variants. The number of partially shared variants estimated to have arisen at each stage of development, using the adapted LachesisMap method (A). Representative developmental trees of mtDNA variants predicted to have arisen in the endoderm (B), mesoderm (C), and ectoderm (D) germ layers. The number and color of nodes indicate tissue type (Supplementary Fig. S25 for color key). Edges colored purple represent developmental time points where the given variant could have occurred. Terminal nodes colored purple represent samples with the given variant, and internal nodes colored purple represent developmental tissues in which the variant was likely to have been present.