Thread by @Eric_Fauman

I love the idea of this GWAS. The authors estimated the abundance of mtDNA in the blood of @uk_biobank participants by using the intensities of probes mapping to the mito genome

@HaggSara
Juulia Jylhävä
Yunzhang Wang
Kamila Czene & 
Felix Grassmann

pubmed.ncbi.nlm.nih.gov/33385171/

There are many analyses in this paper; I'm (naturally) only focusing on the GWAS which identified 66 lead SNPs in 50 loci.

I think I found 4 likely causal genes not mentioned in the paper including an explanation for the strongest hit and largest beta out there on Chr 19.

As a first step the authors used FUMA to select a functional variant for each lead SNP, picked the closest gene to each SNP and looked for shared themes among those genes. This found 3 plausible themes:
1) immune related,
2) cancer & cell cycle and
3) mitochondrial function

A quick segue before we get to the new mitochondrial genes.

As revealed above, many of the genes in this GWAS relate to immune cell abundance, I assume because WBCs have mitochondria while RBCs do not.

Relevant genes include MYB near HBS1L for RBC count & JMJD1C for WBCs

Ok - let's focus on the mitochondrial signals.
When I explore the closest genes I see convergence for

REACTOME_TRANSCRIPTIONAL_ACTIVATION_OF_MITOCHONDRIAL_BIOGENESIS: MEF2C, TFAM, NR1D1

and for

GOCC: mitochondrial nucleoid: POLG, TERT, TFAM

This reveals the first new gene: NR1D1.

NR1D1 encodes Rev-erbα, a nuclear hormone receptor involved in regulation of cell cycle.
A degradation resistant mutant of NR1D1 results in increased mitochondrial abundance.
ncbi.nlm.nih.gov/pmc/articles/P…

Not sure why this gene wasn't mentioned

For step 2 we use the fact that the closest gene is usually correct, so we can use the themes from the closest genes to find candidate causal genes that are further from the lead SNP.

~50 genes are included in the reactome or GO sets above.

Are any of these near SNPs?

Just 20kb from rs741735 we find SIRT3. SIRT3 is a mitochondrial-specific deacetylase. SIRT3 works with FOXO3 to regulate mitochondrial biogenesis:

A novel AMPK-dependent FoxO3A-SIRT3 intramitochondrial complex sensing glucose levels

pubmed.ncbi.nlm.nih.gov/23283301/

At 100kb from rs1408343 is my favorite gene name of the day: TWNK

TWNK encodes a mtDNA-specific helicase.

Human mitochondrial DNA deletions associated with mutations in the gene encoding Twinkle, a phage T7 gene 4-like protein localized in mitochondria

pubmed.ncbi.nlm.nih.gov/11431692/

Here's the explanation for the naming of TWNK:

Because of the unusual localization pattern, reminiscent of twinkling stars, we designate the full-length protein Twinkle
(for *T*7 gp4-like protein *w*ith *i*ntramitochondrial *n*ucleoid *l*ocalization).

pubmed.ncbi.nlm.nih.gov/11431692/

So we still have that signal with a p-value of 3e-62 on Chr19, locus 43 which actually comprises 4 independent variants covering a 200 kb window.
rs806709 is 74kb from LONP1 (rs566777150, p=1e-18 is only 20kb from LONP1).
Why is LONP1 annotated with GOCC: mitochondrial nucleoid?

LONP1 encodes a mitochondrial protease which regulates numerous mitochondrial functions:

1) it degrades damaged proteins
2) it degrades specific oxphos components
3) it regulates mtDNA replication by degrading the phosphorylated form of TFAM

pubmed.ncbi.nlm.nih.gov/26363553/

Interestingly that lead GWAS SNP is in high LD (r2 0.8-0.9) to a missense variant in LONP1: Arg241Gln. No information on whether that variant impacts activity of the protease.
ldlink.nci.nih.gov/?var=rs806709&…

LONP1 is responsible for a severe Mendelian disease, CODAS syndrome:

Cerebral
Ocular
Dental
Auricular
Skeletal
anomalies syndrome

The disease features "abnormal mitochondrial inner-membrane topology"

omim.org/entry/600373

So that's it - 4 unannotated loci with likely causal genes related to regulated of mitochondrial biogenesis:

NR1D1 - a cell cycle transcription factor
TWNK - a mtDNA helicase
SIRT3 - a mt histone deacetylase
LONP1 - a mt protease

Each could be worthy of further exploration

As a structural biologist, I'm always drawn to the protein structures of the causal genes.

The image I used for LONP1 is from electron microscopy on the E coli homolog, Lon. The bacterial protein exists as dodecamer

pubmed.ncbi.nlm.nih.gov/23674680/

Epilog on the Lon structure - electron microscopy leads to the purple image.

"Distinct quaternary structures of the AAA+ Lon protease control substrate degradation"

Ellen F Vieux, Matthew L Wohlever, James Z Chen, Robert T Sauer, Tania A Baker

pubmed.ncbi.nlm.nih.gov/23674680/