Our analysis of right heart structure and function is now published in
@NatureGenet
.
We used deep learning to measure the right atrium, the right ventricle, and the pulmonary root & artery in the
@uk_biobank
.
Paper is here:
Welcome
@jpirruccello
! Dr. Pirruccello joins
@UCSF
in Oct from Mass. General Hosp. He brings with him an impressive research program and
@NIH
funding in cardiovascular genetics and will help lead the cardiovascular cohort in 3D health
@UCSF_BCHSI
.
Bigger aortas are more prone to tear, but are usually asymptomatic until then. So, it would be nice to know the size of one's ascending aorta.
πDid we measure everyone's ascending aorta? No.
But we did build a simple clinical estimator, published today:
Simple idea:
When conducting rare-variant burden tests to discover associations between genes and continuous phenotypes,
you can get better results by adjusting for a polygenic score for the same phenotype.
A 𧡠π
At Mass General Brigham, we are coming together to enhance clinical care, research, and education for the patients and communities we serve. Now, we are taking the next transformative step in that journey, beginning a process to combine our academic medical center departments.
Excited to share work done with
@KrishnaAragam
, where we discovered common genetic variants associated with the structure and function of the heart. These common variants, found in healthy people, also influence risk for heart muscle diseases.
Our study looking at genetics of ascending and descending thoracic aortic strain and distensibility in 40,000
@uk_biobank
participants is now out in
@JACCJournals
(full article available via this link).
Strain doesnβt just recapitulate diameter.
We derived aortic valve measurements in ~50k UK Biobank participants.
Even among healthy people, aortic valve function exists on a continuum, influencing risk of clinically significant aortic stenosis and heart failure.
(Images reproduced by kind permission of the UK Biobank)
Valvular heart disease is common across the globe. We have a good understanding when to treat disease such as severe aortic stenosis. But we know little about normal variation of valvular function and where disease begins. So we looked at the UK biobank.
Brief walkthrough to give background on work with
@JenHoCardiology
,
@patrick_ellinor
&
@steven_lubitz
looking at genetic factors associated w/ variation in right heart structure & fxn. This is work in progress and feedback would be welcome!
First, a methods-focused thread.
We have claimed, for Electoral Vote purposes, the Commonwealth of Pennsylvania (which wonβt allow legal observers) the State of Georgia, and the State of North Carolina, each one of which has a BIG Trump lead. Additionally, we hereby claim the State of Michigan if, in fact,.....
@johnlimouze
Lots of people who donβt practice medicine seem to be really enthusiastic about giving, to those of us who do, the ability to deny care to the unvaccinated.
We donβt want this!
I would like these people to pick a different venue for their political revenge fantasies.
In 2019, a Mendelian randomization paper in
@NEJM
suggested that ACLY is a good drug target for LDL.
I am still not convinced that it tells us anything about ACLY.
@rwyeh
@EricSecemskyMD
@HeartBobH
Conclusions:
1. The entirety of the protective effect of the exome (and cardiac MRI) is attributable to immortal time bias from cardiac MRI recipients.
2. This can be completely controlled for by deleting your immortal time.
3. This must be accounted for in the UK Biobank.
@justsaysinmice
This is a great account.
Worth noting that nobody calls lab mice "patients," so just reading the first... 4 paragraphs... would be extraordinarily misleading.
The Massachusetts COVID dashboard doesnβt graph ICU admissions, but it does provide the raw data so you can plot it yourself. From todayβs data release:
Updated LV genetics analysis:
- Now in 36,000 people & peer reviewed
- More loci, TWAS as a secondary prioritization tool at each locus
- Stronger PRS effect on DCM (1.58x HR per SD)
- Replication (as best we could)
Available as an open-access pub at
Excited to share work done with
@KrishnaAragam
, where we discovered common genetic variants associated with the structure and function of the heart. These common variants, found in healthy people, also influence risk for heart muscle diseases.
Announcing our new GWAS of the number "Ο".
Some highlights (π§΅π):
1. Numbers are not heritable(???)
2. We couldn't even finish writing down pi because it just kind of... kept going? (See our 400 page supplement for our best effort.)
Stay tuned for our upcoming "π" GWAS.
Professor
@UREssien
begins
@BrighamWomens
Medicine Grand Rounds with the ongoing history of racism and eugenics in medicine.
(And of course I have another meeting that starts at 12:15 but I'll catch you on the recording.)
Status: Updating Harvard-template CV for grant applications.
Frustration level: High.
Gratitude for my
@MGHHeartHealth
program director who is helping me with this minutiae: Much higher.
Thank you,
@ddefariayeh
!
Breathtakingly rapid development of effective vaccines.
Dramatically rising COVID case count driven by politics and structural inequity.
A self-encapsulated study of how the US spends the most on healthcare and develops great therapies, but has poor health outcomes.
Congratulations
@MahanNekoui
, Massachusetts General Hospital and Broad Institute, on winning our Early Career Researcher 2022 competition! Tune in to hear more about the fascinating research his team are doing using
#UKBiobank
data here:
#ECR
#UKBSC22
@daniela_witten
What a coincidence; I just got hired by twitter today. I am now in charge of the feature that adds lil crypto scam posts* below every famous personβs tweets.
* = Send me one prime number and Iβll send you two back.
@jeremyphoward
I wrote to
@Scott_Wiener
(my state senator and the bill's author) encouraging him to abandon this misguided bill. It cannot help and it will do harm.
We should regulate actions, whether those actions are informed by large models, small models, heuristics, coin flips, or whims.
Any reasonably led government would have sourced a sufficient supply of masks so that at this point - 6 months into a pandemic - hospital staff *and people in the community* could wear N95s whenever they wanted.
#Masks4All
I worked with some unionized fellows this week in the UCSF CCU and they were extremely good at their jobs.
xoxohth I will not be taking questions at this time
One thing I like about living in BOS and working in Cambridge is the variable reward payout from riding the
@MBTA
Will there be trains every 2 minutes? Or every 20? The uncertainty and occasional big wins keep me coming back for more.
At least 3 times during fellowship people (not at MGH) confused me and
@kparuchuri
(presumably because we have unusual last names that start with 'P').
Now we are giving them a reason. H/t
@pnatarajanmd
The intersection of:
people who would be good at an admin job,
people who want to do an admin job,
and people who would be selected for an admin job
is often an empty set.
I donβt understand why those of us who work in a hospital *arenβt* required to take periodic asymptomatic COVID tests.
Meanwhile, we are continuously taking healthstream tests.
The last time I left
@skathire
and
@kiranmusunuru
it was to send off a company that I had built.
Today, Sek and Kiran are launching their own. Turnabout feels like fair play here!
Now, have a blast, develop a great platform, and eradicate coronary disease.
This is an interesting study but it doesnβt, by itself, establish causality. It is an important flagpole and it will spur good follow-up work.
For completeness, it may be that the same thing that causes microplastic deposition also causes ischemic disease, making this a marker.
Patients with evidence of microplastics and nanoplastics in carotid artery plaque, as compared with patients without, had a greater risk of adverse cardiovascular events at 34 months of follow-up. Read the full article:
> It's "unconscionable ... an experiment on human beings that is not morally or ethically defensible," said Dr.
@kiranmusunuru
, a University of Pennsylvania gene editing expert and editor of a genetics journal.
Well said.
@DrSarahWakeman
@choo_ek
@AnupamBJena
I donβt always mansplain topics outside of my expertise, but when I do, I make sure to do it at a forum about the problems women face.
Shout out to my friends who still get to do HealthStream training today.
Sadly, I spent my year's HealthStream supply a few days ago and won't get to take a hit of that sweet, sweet pretest failure until December 2021.
Kind of obsessed with the chaotic nature of the Epic βSent Messagesβ folder. No alphabetical ordering. Dozens of categories. I dare you to try to find what you were looking for.
Good Morning
#MGH
! Today's Coffees: Caramel Nut Biscotti & Winter Wonderland. Muffins: Corn, Blueberry and Almond Joy, and Be Fit Apple Streusel Bran. Tea Bread: Cranberry Nut.
@spreekaway
@paulg
The 5000 trader limit that you pointed out and the $850 max limit make PredictIt an inefficient market. Basically free money on the table as long as the gain is >5% (the PredictIt withdrawal fee).
Most of the biobank-based disease analyses that we are able to do right now are based on billing codes.
Developing algorithms that reliably (or even reliably-unreliably) pick up diseases that don't have ICD codes would be a boon.