Profile avatar
ekernf01.bsky.social
Statistician and computational biologist; uw alum; jhu student. He/him. http://ekernf01.github.io
266 posts 1,213 followers 391 following
Regular Contributor
Active Commenter
Posts 24 Comments 25

For me, it's roughly day 100 of being pleasantly surprised that a position-weight matrix can capture a complex-seeming distribution over zillions of possible antibodies... ... and then bust out of the screen and be manufactured in real life.

submitted 1 day ago • 0 comments

Apparently, transcriptomics is not the only field struggling with overly optimistic results from new deep learning techniques. www.understandingai.org/p/i-got-fool...

submitted 3 days ago • 1 comment

Why do you need a PhD in computer science to find an antigen expressed in one cell type but not the rest? www.indeed.com/viewjob?jk=5... The work looks important and it would be a privilege to contribute. But, see question above.

submitted 4 days ago • 1 comment

I'm defending my PhD next Friday, May 23!(!!!!). I'll be highlighting our work looking at aneuploidy in early human development. If you're interested I'd love to have you join via Zoom (DM me for info) or on the Homewood campus!

submitted 7 days ago • 2 comments

Easily the most important thing happening next week. Come and watch my friend Sara defend her PhD!

submitted 7 days ago • 0 comments

I phoned the Baltimore DOT to thank them for putting traffic calming near my home. They said "We don't get many of these. We only hear it when it's bad."

submitted 8 days ago • 1 comment

Cellular plasticity always has fewer rules than I expect. Blood cells can crawl out of your vessels and start laying down cartilage to fix burns. doi.org/10.1097%2F01...

submitted 8 days ago • 1 comment

There are, right now, very deliberate and carefully planned philanthropy projects aiming to unlock new discoveries across many fields of science. www.gap-map.org

submitted 9 days ago • 0 comments

Multiplexed perturbation screens facilitate a key prerequisite of "virtual cell" efforts: getting lots of perturbation data with detailed outputs. However, they present unique experimental design obstacles. 🧵

submitted 9 days ago • 1 comment

🚨 I just released the alpha-version of my AI in Genomics textbook! It covers (so far!) sequence language models & deep learning for DNA/RNA/proteins. Check it out: biobook.michel.science

submitted 11 days ago • 1 comment

We say a statistic U is complete for a family theta when: E[v(U)|theta] = 0 for all theta ==> v(U)=0 "Complete" is the worst possible name. I propose a better name: "parsimonious".🧵

submitted 10 days ago • 1 comment

This looks like a great opportunity -- focus on adding statistical rigor to clinical ML/AI. www.jeanfeng.com/joining

submitted 10 days ago • 0 comments

I wrote a guest post on a good friend and colleague's excellent newsletter, the Connected Ideas Project. It's about moving from academia to industry in biotechnology doi.org/10.59350/z2s...

submitted 11 days ago • 0 comments

I would like to pay someone to help make my personal website look a little better. It's Jekyll. If interested, please DM for details.

submitted 13 days ago • 0 comments

Nominative determinism strikes again.

submitted 14 days ago • 0 comments

Back around 2020, I thought we'd never have the resolution to map enhancer-gene contacts en masse with DNA looping assays. But region-capture micro-C is now making this possible.

submitted 15 days ago • 0 comments

👇

submitted 15 days ago • 0 comments

Big new study of eQTL's whose effect depends on drug exposures finds that drug-dependent eQTL's are more like disease risk loci in general (context, constraint) and sometimes overlap disease loci when plain eQTL's do not.

submitted 16 days ago • 0 comments

New blog post: pitfalls and open questions in GRN inference, ft. Audrey II. ekernf01.github.io/guiding-ques...

submitted 17 days ago • 0 comments

Implement a 100% tariff and ruin a movie Close Encounters of the 6th Kind

submitted 18 days ago • 0 comments

Biotech is discovering the awesome power of fractional factorial designs (or similar randomized designs) and I'm all for it. en.wikipedia.org/wiki/Fractio... www.nature.com/articles/s41...

submitted 19 days ago • 1 comment

Tempus: a company that I did not know existed, but in fact publishes REAMS of successful-seeming cancer diagnostics research. www.tempus.com/publications/

submitted 20 days ago • 0 comments

Verge: a company to check out for people interested in gene regulatory networks and perturbation profiling of human cells. www.vergegenomics.com/approach

submitted 21 days ago • 0 comments

I did* a deep dive on a fancy MCMC method for reaction rate parameter inference: - some molecules not observed - measurements intermittent, not continuous - system is stochastic (low molecule counts) - all reactions known github.com/ekernf01/pre... *I wrote this 10 years ago (!?)

submitted 21 days ago • 0 comments