777 studies, 745 with summary statistics - the @genesandhealth.bsky.social WES study is now in the GWAS Catalog, aiming to improve health outcomes for South Asian populations. Go explore 👇
www.ebi.ac.uk/gwas/publica...
It's not just the burden on researchers to do the reporting. Someone, somewhere must be collating, evaluating all these reports, surely? Then this information must be fed to a committee, who in turn decide whether we are doing as we promised. That's a lot of hours. What is the cost-benefit ratio?
Some genuinely good news!
Completely agree!
Also, to my co authors, this is why my first drafts look like they do 🫣
A cycle way in Cambridge which is red tarmac with loose gravel on top. To the left you can see a bit of the road (black tarmac) which is breaking up, leading to the gravel on the cycle way.
In Cambridge the decaying road surface turns the bespoke cycleway into a gravel track. The holes next to the cycle way are now so deep that buses routinely straddle them by putting one set of wheels on the cycleway, which is dangerous and will lead to degradation of the cycle way in the long run.
Just discovered ggmagnify hughjonesd.github.io/ggmagnify/re.... love it! But quite probable that I will overuse slightly creating figures for my next talk. You have been warned.
What I most want employers to understand is that writing code and performing meaningful data analysis are not the exact same skillset. #rstats
Well, good luck with your reply
Always hard, and the grant body knew in advance the timings, surely?
A WHOLE CIVILIZATION WILL DIE TONIGHT My son needs lunch, and I have to put his backpack together, but a whole civilization will die tonight, so I'm wondering if they've closed their schools. Like, a snow day, maybe, except instead of snow it's "keep your children home so if you die, you die together" — instead of "well open back up once the plows have cleared" it's "we don't know if we'll be here tomorrow, hold your babies tight." It's just "talk" I'm told, which I've been told before. "It's how the president makes his deals." But I've never heard anyone talk about other human beings this way, and I'm not certain I can look my son in the eyes if we all agree to stomach it one more time. A civilization will die tonight, but as I zip up his backpack and kiss him off to school I think: if this is what we call leadership then I'm not entirely sure ours isn't already dead. @michaelfdubois Mukad A QuBoy @michacifdubois
Brutal.
notes on my benchmark of my model vs your model with my metric
Thank you so much for this summary. Can I ask where you saw this detail on those who applied in September 2025? Is this just for reapplications?
Agree. Simplify where possible. But there can be beauty in long form insults.
📢We’re hiring!
3 Research Associates to join Professor Arthur Kaser's lab at #CITIID to investigate FAMIN, a core enzyme we discovered that prevents inflammation and autoimmunity.
➡️https://buff.ly/Gtbb6ss
➡️https://buff.ly/W4cr9xC
➡️https://buff.ly/kDv7VdI
Apply by 9 April.
To homeopathic concentrations, I hope
That's how I think of it. Agree this shouldn't happen in a randomised trial. This is a bad choice of example.
Is this to describe the candidate, myself or both? 🤣
"Join us, together we can cling to the sides of the life raft we are building whilst traversing some challenging rapids to reach bright new scientific horizons."
A thought about job adverts. I assume seeking "exceptional" individuals selects for specific character traits (eg self confidence). My goal is to get candidate with best potential, even if they don't yet think of themselves that way. What words do you use to encourage the widest pool of applicants?
Omg. Floodlit indeed!
Clearly we both need the lighting people to find a middle ground
Do we live in the same Cambridge? Here the lights are so dim you can't see the pavement as you walk between them.
Thanks! Very clear explanation. Hadn't seen is_vector() before. Useful in this case, but confusing to have two functions with such similar names that will return different answers.
glue::as_glue() documentation says result is a vector with S3 class "glue". But
is.vector(glue("thing"))
returns FALSE
Can anyone explain?
#rstats conundrum
R often produces PNG with a transparent background which appears black if you use bluesky in dark mode. From memory, this is resolved by adding ' bg="white" '.
Pleased they are doing this for basic anti corruption reasons. But also, if they were actually developing a little political nous against reform? A good thing too.
Interesting. I may try and find some to test myself. Thanks so much for sharing!
The plate looks gorgeous. The lions mane itself... looks not very vegetarian 🤣 how does it taste???
The UK could do this too. We should.
Brian Ripley only posted about CRAN policy once in 2009???
Fantastic resource though, thank you!
The kind of drawings only a mother would love!