New preprint from my group led by Yuchang Wu Zhongxuan(John) Sun. Almost every recent Alzheimer's GWAS combines clinical AD diagnosis and proxy phenotypes based on parental health survey. We report major limitations in the current practices of GWAS-by-proxy (GWAX) biorxiv.org/content/10.110…

This! I came to very similar conclusions when looking at by-proxy vs clinical AD phenotypes during my PhD. Specificity matters.

I still stand by this one too

Did you spot us on @bbcbreakfast? Today we hit 1 million people signed up to #OurFutureHealth, all thanks to our brilliant volunteers across the UK. Watch on iPlayer from 2hr30min to hear more about this achievement: bit.ly/49lBfWW

Mel Bartley Here's the mind blowing thing about selection bias. If you end up with a biased sample of healthy wealthy people, your results aren't even generalizable to a population of healthy wealthy people!

Finally Simpson’s paradox in a way even I can follow …

Hard agree

If I was fired then immediately hired by Microsoft, only to then get my original job back b/c majority of the people under my leadership threatened to quit. Then the people that fired me were then fired. All in less than a week…you wouldn’t be able to tell me ANYTHING 😂

📣Our 🆕 paper Causal Inference is Not Just a Statistics Problem is out! Malco`lm()` Barrett (Now on Bluesky and Mastodon), King Ranch @[email protected], and I show that you can have 4 data sets with identical summary stats & visuals but very different data generating mechanisms-statistics alone can't tell you what to adjust for!

2024: publish your analysis code or it did not happen

Pleased to see “Common infections and neuroimaging markers of dementia in three UK cohort studies” is now published in Alzheimer’s and Dementia! A fantastic collaborative effort during my postdoc MRC LHA @ UCL - thanks to PI Dylan Williams and all coauthors ⭐️ alz-journals.onlinelibrary.wiley.com/doi/10.1002/al…

Product management is actually pretty easy. You just need to know design, finance, elements of engineering, psychology, negotiation, project management, speaking, writing, marketing, storytelling, analytics, never stop being curious, and want to learn more every single day.

This is a brilliant talk by Peter Tennant on collider bias, why sample selection matters, and why transparency about the issue being present is key to ensure good statistical and reporting practices.

Everybody needs a small group of highly driven, ambitious, curious friends, that they can casually meet up every week, to talk about how they can reach their respective goals more efficiently, how they can contribute to society more meaningfully, how they can simply be better.

Can confirm - wonderful team. For anyone looking for a new role in clinical trial statistics 👇🏻

📣 👇🏻 Researcher community 👇🏻 📣 Don't miss this opportunity to be amongst the first to make use of and help shape the Our Future Health resource. Our Trusted Research Environment already has data available for >700,000 participants, and is growing every quarter. Apply below.

Dr Oliver Pain & Prof Dr Cathryn Lewis have launched GenoPred - a tool that allows researchers to easily calculate #polygenic scores to uncover individuals' genetic risk for certain health-related outcomes 🧬 🔗 👉buff.ly/4c1hjJu

Congratulations to everyone accepted into our new Early Adopter Programme for researchers! They’re among the first to access #OurFutureHealth and will advocate for our programme in the wider research community. Read the full list of Early Adopters here: research.ourfuturehealth.org.uk/early-adopters…