@timfrayling
Human genetics, type 2 diabetes and obesity in today's environment. Cycling uphill. Team science. New shoots in Geneva & France, long roots in the UK & Exeter. https://www.unige.ch/medecine/gede/en/research-groups/timothy-frayling
Learn about working securely with data from 10% of the Swiss population :
📚 The SPHN Federated Clinical Routine Dataset:
🗣 Referenten: Dr. Sabine Österle
📅 Wann: Dienstag, 10. März 2026, um 16:45 Uhr
📍 Wo: Online via Zoom: bit.ly/4e9p3Ka
#SPHN
Amazing. UK leading the way . Finland in silver medal position ? and I predict Switzerland could join them on the podium at some point over the next year or two
& here's a very good example: "The proposed data-driven clusters differ in diabetes progression and treatment response, but models that are based on simple continuous clinical features are more useful to stratify patients" www.sciencedirect.com/science/arti...
yep. seems like we need to try, but also that most components of disease are on a continuous scale. Put people into groups based on those scales and they will look different, but are the separations real ? doesn't rule out the importance of stratification to get more bang for intervention buck
Worth keeping an eye on Spain who seem to be going down a different, more evidenced based, route : www.theguardian.com/world/2026/j...
Congratulations @laferrat.bsky.social & team.
slightly random reminder of this paper from 2019 from Exeter colleagues that shows how heterogeneity of disease, in this case diabetes, is best dealt with by using continuous clinical variables not cluster assignment: www.sciencedirect.com/science/arti...
excellent opportunity with SPHN , making health data from > 10% of the Swiss population safely reusable for research apply.refline.ch/499599/0353/...
🧬 New preprint alert! After years of collaborative work across 52 datasets we are presenting eQTLGen phase 2: a genome-wide eQTL meta-analysis covering 43,301 blood samples: www.medrxiv.org/content/10.6... (1/8)
Using genetics to understand multi morbidity, including neat R package to "subtract" effects of critical risk factors such as obesity : www.nature.com/articles/s43...
great work & just as interesting is use of this approach to estimate the shared genetics between pairs of conditions whilst "subtracting" the effects of obesity. Interesting that T2D-Osteoarthritis genetic link estimated as 100% about obesity, other pairs clearly have strong non obesity components
The Exeter team seem to be the Alex Honnold of human genetics. Most sane people on seeing the challenge of meta analysing >1billion variants & 10,000s aggregates across ~100% of the genome would opt for the safety rope of the exome. Not these guys.
ten percent of the Swiss population: Medical meta data update: www.linkedin.com/posts/swiss-...
Learning more about the brilliance of the medical record data at the University Hospital Geneva. This plot, from pubmed.ncbi.nlm.nih.gov/34643542/ shows how medical informatics researchers have introduced , & clinicians passively adopted, a set list of phrases, reducing new free text.
Interesting development in ”non-personalised Medicine” from UK:
the government will reduce the current lower threshold at which Soft drink industry levy applies from 5g of total sugars per 100ml to 4.5g
& will remove the current exemption for milk-based drinks with added sugar
Rarely in the entire history of science, has QC been a topic of such passion, importance and impact. If only the French and Americans had adopted such rigor when they messed up the design of that multi billion $ telescope because one was using the metric system, the other the imperial system.
the last one was a nice review from Hanieh Yaghootkar, this is the partitioned PRS paper: www.nature.com/articles/s41...
Now a nice series of papers using genetics to "uncouple" excess adiposity from its adverse metabolic effects. Here for metabolic traits www.nature.com/articles/s42..., here including non metabolic traits elifesciences.org/articles/72452 & here using partitioned PRS www.nature.com/articles/s41...
Nearly two years in for our new common disease genetics and genomics team at the University of Geneva. Already lots to be proud of. Here is Lauric talking about his award from the local Diabetes charity. fr.linkedin.com/posts/fondat...
I’m learning more and more about the extremely impressive Swiss personalised Health Network and the trusted research environment infrastructure that enables researchers to safely access & Analyse data from multiple hospitals in Switzerland www.sib.swiss/news/sphn-is...
Congratulations to our Liza Darrous for winning the women in data science award last week: tinyurl.com/lizaprize
For more details about the Swiss personalised Health Network and the safe way trained researchers can work with data from multiple hospitals : sphn.ch/network/proj...
thanks to @laraedw001.bsky.social for highlighting this micro.green-park.co.uk/hdrs/. I suggest whoever gets this position, pops over to Switzerland as the SPHN have achieved great things with consented data from 10% of the population available for study a few months after application & in a TRE
When you say "nobody" do you mean "nobody outside of your health care providers" ? your hospital will use your data all the time to help audit and improve their services ?
smart work aged 8. but possibly redundant for 17 years time when birth rates in europe may be 0.1 per couple. Maybe aged 85 : lose large amount of money to pay for private social care ?
Oh dear. Sad and presumably another lead statistician to be fired ? www.bbc.com/news/live/c5...
thanks Laura. Yes I think the 1st slide on PRS talks needs to be "do not compare PRS to clinical genetics, compare to the common, equally bad & equally good , stuff your doctor measures all the time to stratify you - LDLC, blood pressure, family history, ancestry and BMI" (to add to your age & sex)
yes . I guess my slight air of frustration is that we have known for while that PRS are complementary to other risk factors e.g. journals.plos.org/plosmedicine...
the first one does not seem to be the right question.
