henhaobio
Sign in

NatureGenetic & Genomic MedicinePeer-reviewed journal article

Nuclear genetic control of mtDNA copy number and heteroplasmy in humans

Rahul GuptaiD, Masahiro KanaiiD, Timothy DurhamiD, Kristin TsuoiD, Jason G. McCoyiD, Anna V. KotrysiD, Wei ZhouiD, Patrick Francis ChinneryiD, Konrad J. KarczewskiiD, Sarah E. CalvoiD, Benjamin M. NealeiD, Vamsi Krishna MoothaiD

PDFDOIPreprintCodeData
In-depth full-text digest available · Jump to digest ↓

Abstract

Human mitochondria contain a high copy number, maternally transmitted genome (mtDNA) that encodes 13 proteins required for oxidative phosphorylation. Heteroplasmy arises when multiple mtDNA variants co-exist in an individual and can exhibit complex dynamics in disease and in aging. As all proteins involved in mtDNA replication and maintenance are nuclear-encoded, heteroplasmy levels can, in principle, be under nuclear genetic control, however this has never been shown in humans. Here, we develop algorithms to quantify mtDNA copy number (mtCN) and heteroplasmy levels using blood-derived whole genome sequences from 274,832 individuals of diverse ancestry and perform GWAS to identify nuclear loci controlling these traits. After careful correction for blood cell composition, we observe that mtCN declines linearly with age and is associated with 92 independent nuclear genetic loci. We find that nearly every individual carries heteroplasmic variants that obey two key patterns: (1) heteroplasmic single nucleotide variants are somatic mutations that accumulate sharply after age 70, while (2) heteroplasmic indels are maternally transmitted as mtDNA mixtures with resulting levels influenced by 42 independent nuclear loci involved in mtDNA replication, maintenance, and novel pathways. These nuclear loci do not appear to act by mtDNA mutagenesis, but rather, likely act by conferring a replicative advantage to specific mtDNA molecules. As an illustrative example, the most common heteroplasmy we identify is a length variant carried by >50% of humans at position m.302 within a G-quadruplex known to serve as a replication switch. We find that this heteroplasmic variant exerts cis-acting genetic control over mtDNA abundance and is itself under trans-acting genetic control of nuclear loci encoding protein components of this regulatory switch. Our study showcases how nuclear haplotype can privilege the replication of specific mtDNA molecules to shape mtCN and heteroplasmy dynamics in the human population.

In-depth summary

Claude Opus 5.5 · xhigh effort · 8 Nov 2026 · from the full text of version 1 (19 Jan 2023) · CC BY 4.0

Blood whole genomes from 274,832 people yield 92 independent nuclear signals for mtDNA copy number and 42 for inherited heteroplasmy levels, implicating mtDNA replication machinery.

Although every protein that replicates and maintains mtDNA is nuclear-encoded, nuclear genetic control of human heteroplasmy had never been shown, and many earlier blood mtCN associations appear confounded by blood cell composition. Biobank-scale mtDNA calling with careful confounder correction maps these nuclear loci, separates inherited from somatic heteroplasmy, and reassesses reported mtCN-disease links.

Key results

  1. ShownAfter correction for blood composition, mtDNA copy number (mtCN) falls linearly with age and is under polygenic nuclear control. Implicated genes work in the nucleoid/replisome, nucleotide metabolism and mitochondrial fusion; other loci, such as PNP/APEX1, MCAT and NDUFV3, had no previous link to mtCN.

    mtCNcorr in 178,134 UKB subjects (mean 61.66 copies per nuclear genome) fell ∼2% per decade; 92 LD-independent signals at 46 loci, SNP-heritability ∼4%; haplogroup explained < 0.5% of variance; 33 loci had variants with PIP >= 0.1, 11 with protein-altering credible-set variants; APEX1 missense at PIP > 0.9. · Determinants of mtDNA copy number variation

  2. ShownBlood cell composition confounds many reported links between low blood mtCN and common disease, as well as several blood- and inflammation-related mtCN GWAS loci. Mendelian randomization suggests that cell composition drives raw mtCN rather than the reverse.

    Blood composition explained R^2 ∼23% of raw mtCN; mtCNcorr showed no link to type 2 diabetes, myocardial infarction, stroke, hypertension or dementia (recovered with mtCNraw); of 24 other diseases only osteoarthritis had lower mtCN; 70 of 96 prior loci replicated at p < 5*10^-5 (37 genome-wide), 12 more (e.g., HLA) only with mtCNraw; TFAM signal rose ∼40 orders of magnitude. · Determinants of mtDNA copy number variation; Blood cell composition confounds prior genetic…

  3. ShownHeteroplasmy splits into two classes. Single-nucleotide variant (SNV) heteroplasmies behave like somatic, transition-rich mutations that build up late in life. Indel heteroplasmies, found next to non-coding poly-C tracts, stay stable with age and pass down the maternal line as mixtures.

    In AoU (ages 20-90 vs 40-70 in UKB), SNV heteroplasmy counts rose with age, most sharply after 70, while indel counts did not; in UKB mother-offspring and sibling pairs, nearly all indel levels were shared but most SNVs were not; no paternal transmission; most people carried 0-1 heteroplasmic SNVs and 0-2 indels; indels present in >60% of individuals. · Spectrum of mtDNA sequene variation across…; mtDNA SNVs and indels exhibit distinct modes of…; DISCUSSION

  4. ShownLevels of common heteroplasmies share a nuclear genetic architecture centred on mtDNA replication and nucleotide-balance genes. The loci appear to shift inherited heteroplasmic fractions rather than cause new mutations, and effect sizes replicate in All of Us (AoU).

    42 LD-independent associations across 39 heteroplasmies in UKB, with nine of 20 loci shared by >1 heteroplasmy; PNP intronic variant PIP 1 (eQTL colocalization PIP 1), DGUOK p.Gln170Arg PIP 0.99, POLG2 p.Gly416Ala PIP 1; case/control recoding gave little signal; AoU replication in 96,698 people, effect-size R^2 = 0.79. · Nuclear genome GWAS for mtDNA heteroplasmy; DISCUSSION

  5. ShownThe most common heteroplasmy is a chrM:302 length variant in the conserved sequence block II (CSBII) G-quadruplex replication switch, carried by >50% of people. It occurs as inherited mixtures even within single cells, associates in cis with lower mtCN, and is shifted in trans by replisome-related nuclear loci.

    Mixtures seen in most of 156,885 UKB individuals and in 171 single cells from one person (mtscATAC-seq); longer alleles tied to lower mtCNcorr with effect comparable to the TFAM locus (PIP ∼ 1); TFAM 5' UTR and LONP1 missense variants had PIP ∼ 1 for mtCN and PIP > 0.3 for chrM:302:A,AC; SSBP1 variants favored chrM:302:A,ACC. · Pervasive length variation in CSBII across…

Methods, data and reproducibility

mtSwirl, a WDL pipeline, calls mtDNA variants and copy number from blood whole-genome sequencing against per-sample consensus references to limit mis-mapping from nuclear mtDNA-like segments (NUMTs); covariate-corrected traits undergo ancestry-stratified SAIGE GWAS, cross-ancestry meta-analysis, fine-mapping, eQTL colocalization and All of Us replication.

  • UK Biobank blood (buffy coat) whole-genome sequences; primary cohort for mtCN, heteroplasmy and coverage-trait GWAS plus sibling/parent-offspring transmission · human · n = ∼200,000 with WGS; 178,134 in mtCNcorr analysis
  • All of Us whole-blood whole-genome sequences (Controlled Tier v6, Q2 2022 release), used for age analyses and heteroplasmy GWAS replication · human · n = 98,590 sequenced; 96,698 in replication
  • gnomAD v3.1 subset enriched for diverse haplogroups (including L haplogroups) to prototype mtSwirl against prior gnomAD mtDNA calls · human · n = ∼6300 of 76,156 samples
  • Published blood single-cell mtDNA ATAC-seq (mtscATAC-seq) data re-analyzed for chrM:302 length variants · human · n = 171 cells from one individual
  • Pan UKBB GWAS summary statistics (including neutrophil count), ancestry-specific LD matrices and 29 curated common disease phenotypes · human
  • Fine-mapped cis-eQTL results from GTEx v8 and eQTL Catalogue release 4 for colocalization · human
  • Catalog of reference and polymorphic NUMT intervals (BLASTn on GRCh37 plus published sets) for LD-based contamination checks · human · n = 4,736 intervals
  • Summary statistics from a prior UKB mtCN GWAS, used to test replication of earlier loci · human · n = 96 previously reported loci
Methods & tools
mtSwirl (mtSwirlSingle / mtSwirlMulti), GATK v4.2.6.0 (Mutect2, HaplotypeCaller), Haplogrep and Haplocheck, samtools v1.9 and bcftools v1.16, Hail v0.2.98 and Hail Batch, SAIGE v1.1.5, Plink v1.90 (LD clumping), FINEMAP-inf and SuSiE-inf, Stratified LD-score regression (S-LDSC), Variant Effect Predictor v101, deming v1.4 R package (Deming regression), BedTools v2.29.2
Software & environment
GATK v4.2.6.0, samtools v1.9, bcftools v1.16, Haplocheck 0124, Hail v0.2.98, R v4.2.1, SAIGE v1.1.5, Cromwell v77, Plink v1.90, BedTools v2.29.2, VEP v101; WDL pipelines run on UKB RAP, AoU Workbench and Terra.
Reproducibility gaps
Full UKB and AoU GWAS summary statistics are promised on Zenodo only after peer review; AoU individual-level data cannot be released publicly; UKB data need an access application; mtscATAC-seq data require dbGaP approval.
Limitations (6)
  • Heteroplasmies below 0.05 allele fraction and samples with mtCN under 50 were removed because of elevated NUMT-derived artefacts, so very low-level heteroplasmy and low-copy samples fall outside the analysis. · stated by the authors
  • Tests of the most common heteroplasmies against 29 common diseases found little or no association, but the authors note that sample sizes for these comparisons were limited. · stated by the authors
  • The SSBP1 heteroplasmy locus showed LD R^2 ∼1 with variants in a reference NUMT among UKB Europeans; its support rests on the signal persisting in the AoU African-ancestry group, which has lower NUMT LD. · stated by the authors
  • AoU gave no sample collection date, so age was approximated as 2021 minus birth year. The MID ancestry group was not analyzed because fewer than 400 sequenced samples were assigned to it. · stated by the authors
  • All mtCN and heteroplasmy traits come from blood. Whether the nuclear loci act the same way in other tissues was not tested directly; skeletal muscle, for example, was flagged by S-LDSC enrichment. · noted in this summary
  • The replicative-advantage mechanism is inferred from the weak case/control versus strong case-only GWAS contrast and prior in vitro work; no experiment here tests how the nuclear variants affect chrM:302 replication. · noted in this summary

Read this if you study mtDNA copy number or heteroplasmy or mito-nuclear genetic interactions, or if you use blood-derived mtCN as a biomarker and need to account for confounding by blood cell composition.

Citations
193
FWCI
17.23
Field percentile
top 1%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISWachsmuth '16Gupta '25Liu '21Watson '26Wengert '24Kuiper '25
74% semantic matchEvolutionary Biology
Manja Wachsmuth, Alexander Huebner, Mingkun Li+2 more·PLOS Genetics·Mar 15, 2016