Finally - please do reach out and let's grab a coffee if you happen to be in beautiful Copenhagen! :)

..and, of course, to all my fellow student and friends who were with me all the way - true science is a team sport and I would never have been able to climb so high without all of you.

Above all though, a genuine and heartfelt thanks to my doctoral advisors @caina89.bsky.social and Matthias Heinig for their invaluable guidance, training and countless opportunities!

While leaving, I am deeply greatful to everyone at @pioneercampus.bsky.social for making the last five years such a wonderful adventure. Special and many thanks to everyone at HPC, #MUDS and ‪@jherzstiftung.bsky.social‬ for their continued support and making our research possible.

I will exploring the capabilities of patient level multi-model foundation models to large-scale human data sets to learn the molecular basis of human disease - stayed tuned for exciting news on this front! :)

Thank you, very encouraging indeed!

Absolutely! Any chance of a reprint coming?

Thanks so much @anavinuela.bsky.social, glad you like it!! :)

...and here is the big, big thanks to all co-authors in full: @xenophong.bsky.social, Peter Kreitmaier, Pauline Kautz, Leif S. Ludwig, Holger Prokisch, Paolo Casale, Matthias Heinig and
@caina89.bsky.social !!! Thank you for this amazing journey!!! 🤩

Huge shout out to @caina89.bsky.social & Matthias Heinig for their guidance, to all co-authors and acknowledgements for invaluable contributions, to @pioneercampus.bsky.social for support, and to @tuuliel.bsky.social, @alanjhodgkinson.bsky.social & @bhattaarjun2p.bsky.social for stellar prior work!

7/7 Taken together, we provide the first deep dive into mtDNA heteroplasmy across 49 human tissues: exploring its age-related changes, its impact on gene expression, mediation of age-related gene expression changes, and the biological mechanism underlying variations in mtRNA mod.

6/7 Finally, we zoom in on p9 mt-tRNA mod levels to test their impact on mitochondrial transcript processing. We replicate prev assocs between them in blood, test this in all 48 tissues, and further test whether cuts on the 5’ of p9 mods mediate this effect using read-level data.

5/7 There are 10 sites (9 mtDNA het, 1 mtRNA mod) linked to both donor age and mtDNA gene exp. To learn how age, gene exp and het/mod are related, we perform a mediation analysis using partial corr, and find the variations are consistently mediating the effect of age on gene exp.

4/7 Next, we apply the framework to test for relationships with mtDNA encoded poly-A transcript levels, and find 784 effects across tissues (# 576 mtDNA hets; # 208 mtRNA mods). We find that 18 of these are cell type specific, suggesting potentially common effects across cells.

3/7 Since the detection of both variations are affected by RNAseq coverage, we develop a calibrated association testing framework that accounts for coverage effects. Using it, we identify 109 assocs of mtDNA hets and mtRNA mods with donor age, 7 of which are cell type specific.

2/7 We first assess the landscape of both types of variations across tissues: for mtDNA heteroplasmy, we identify those with putatively somatic and inherited origins, and for mtRNA modifications, we find significant enrichments in CNS as compared to other tissues.

1/7 How much does mtDNA heteroplasmy vary between tissues? How much of it is
transcribed to RNA and how much are mtRNAs modified post-transcription? We investigate how these variations change with age, and how they affect mtDNA gene expression in a tissue-specific manner.