reading this intro to lakehouse by databricks. Curious how many biotech companies are using it for data management. My experience of biological data is messy and noisy and not sure how useful the database infrastructure is.

...but as the number of pipelines grows exponentially with the number of steps, in the case of Seurat and Scanpy this will lead to thousands of workflows / hypotheses to validate.

Having a wet lab myself, I can tell you that's not happening. 13/

Scanpy, by default, does not implement clipping and regresses by total counts and percentage of mitochondrial content.

As dana_peer has eloquently pointed out, defaults should only be a starting point. x.com/dana_peer/stat…
So which default should you use, and when? 10/

It's been great to see the positive response of Rahul Satija & Fabian Theis to our preprint biorxiv.org/content/10.110… on Seurat & Scanpy, and their commitment to work to improve transparency of their tools. One immediate benefit will be better practice of PCA in genomics. 1/🧵

Tim Spector MD (Prof) you recent statement that calcium supplements causes heart disease is scaring thousands of my followers who, like me, have had breast cancer and are taking aromatase inhibitors and calcium tablets to prevent osteoporosis

Can you please share the research that backs

Scott Galloway points out the appalling double standard applied to Israel:

“2,200 American servicemen killed at Pearl Harbor. We [America] go on to kill 3.5 million Japanese, including 100,000 in one night. 2,800 Americans in 9/11. We [America] go on to kill 400,000 people in

When some of the self proclaimed world leaders in aging are producing dubious science, you have to say there’s enormous scope to improve standards - but not via this type of organisation

doi.org/10.1016/j.cell…

THE most impactful proponent of Mendelian randomization methods has pumped the breaks, acknowledging his concerns that MR can generate spurious relationships & overstated causal conclusions, especially in nonlinear associations.

TLDR: 🤌obesity still not protective for mortality

This is the person who, right now, is barring the NYPD from dealing with Hamas on campus.

Shame on Columbia University

Lorena Macnaughtan, PhD, MBA Cheshire Elisabeth Bik Jim Bouldin Spotting Jamie Timmons A random cat Captain Science Dror Kolodkin-Gal Please spread the word, Lorena!

In 2018 we published a non-linear Mendelian randomisation (NLMR) study in The Obesity Society and it is now clear the method we applied is deeply flawed, so in October we published a Perspective correcting it. A 🧵on the cautionary tale of NLMR onlinelibrary.wiley.com/doi/epdf/10.10…

I am calling it.

It's time for EVERYONE in the antisemitism task force at Columbia University to resign RIGHT NOW.

Every extra minute that you don't do so is an extra minute of you being an accomplice.

Retweet if you agree.

Earlier today, Columbia University University refused to let me onto campus.

Why? Because they cannot protect my safety as a Jewish professor.

This is 1938.

Hamas continues to hold 133 hostages in Gaza. Several dozen may still be alive, many have been murdered. One is Yagev Buchstab whose sister is a PhD student in the SternLab. Yagev's death was announced on March 23rd by Hamas in a propaganda video. 2/ x.com/lpachter/statu…

I recently came across the use of UMAP to order points by using *1D* UMAP. In journals.plos.org/ploscompbiol/a… we examined how faithful in terms of local and global structure *2D* UMAP is to points in ambient dimension (it isn't). But we didn't look at 1D UMAP. So I took a look now 1/🧵

Russ Hepple Dan Go This suggest change in VO2 max due to exercise training decreases mortality risk

pubmed.ncbi.nlm.nih.gov/31303425/

Dan Go Interestingly, all it seems to take is a high intrinsic VO2max independent of exercise training. No one has tested whether training-induced increase in VO2max can achieve same benefit. pubmed.ncbi.nlm.nih.gov/22867966/

David Grainger These statements are strange and false.

1. Statins' target HMGCoA has overwhelming genetic support.
2. How could a vaccine have human genetic support?
3. GLP1R has variants associated with CAD, fasting glucose and T2D
4. Mutations in the target of Plavix cause bleeding disorders

Maher Arar Learn what genocide means.

More tedious journalism on Vitamin D. This time it’s curing aging. Given most RCTs have found little - perhaps it’s good evidence the aging biomarkers and epigenetic clocks are nonsense? apple.news/AGZFN9XSPS8yJ8…