Jeremy Leipzig
@jermdemo
Followers
2K
Following
8K
Media
370
Statuses
6K
Bioinformatics engr & pop gen product mgmt. O'author, PhD, 4x startup survivor. Pipelines, metadata, genomics, reproducible research. Gen X.
Whitefish, MT
Joined October 2008
Small businesses used to agonize over buying an ad in the yellow pages. Even a quarter page ad would cost the equivalent of $2500 A MONTH in today's dollars.
0
1
1
What startups are focused strictly on reducing friction in getting permissions to random shit for developers?.
0
0
1
Oh please. They reboot brains for the sheer thrill of it.
Here's a must-read deep dive into Bexorg, a biotech restoring dead human brains for better drug testing. A wild one from New Haven, CT from @RLCscienceboss:
0
0
1
In this study, the authors take a panoramic look at how polygenic risk score (PRS) performance for major common diseases has evolved over the past 15 years. While GWAS sample sizes have steadily grown year after year, the gains in PRS accuracy showed strong early momentum but
0
0
0
I remember when Bill Atkinson showed off HyperCard to the Stanford Macintosh Users Group (SMUG). People were literally gasping.
0
0
3
I spent the first 27 years of my career breaking down the most trivial problems for computers to digest and the last month with Claude/Cursor asking "how did you know that's exactly what I wanted?".
1
0
1
Palantir got most of its early funding helping the CIA and NSA with terrorism and surveillance but their real strength is in ontologies. Ontologies come from the crunchiest hippie humanist corners of computer and information science, really library science.
0
0
2
How many hours a weekend do you spend futzing around with weedwacker string?.
0
0
1
Baykal et al "Assessing genomic reproducibility of read alignment tools" - one of the few papers I've seen that study tool robustness.
0
0
0
Learned about an interesting portal spin-off from Fred Hutch called Immediately added it to the "List of cloud genomics companies"
4
4
11
The choice of reference genome is important for reproducibility and reducing artefacts but if the primary endpoint of your experiment only achieves significance with one particular reference you may have a problem with robustness.
Bioinformatics is hard before you even write a single line of code. Here's why. 1/ You haven’t started your DNA-seq analysis. You haven’t aligned a read. And yet you’ve already hit a wall. Which human genome to use?
0
0
2
Am I the only one who doesn't want their smoke detectors talking to each other? I don't need them spreading rumors about fires like they're in grade school. How would I even know who started the lie?.
0
0
2
ok this is sloppy but i still feel bad for people whose figures are ruined by overzealous grammar checks
0
0
1
Gave a talk today and had someone who was actually engaged. I thanked him later and he said he did stand up comedy and remembers being distracted by people talking and texting and shit so he doesn't do that. Something to consider.
0
0
2
"Genomics is not Special. Computational Biologists are reinventing the wheel for big data biology analysis".@jervenbolleman @lh3lh3 @michaelhoffman @antigenomics.
1
0
1
This is probably the single most important thread on biostars, not that the discussion affected Spark adoption directly but you can see a lot of the early signs it wouldn't exactly take bioinformatics by storm.
3
0
5
We're going to see a situation where a lot of life science PIs will come to the bfx core like:."I vibe coded this analysis but it isn't working right. I need you to fix it.".The only appropriate response to this is "Fuck off.".
3
1
3
RT @Jiankui_He: Before you talk to me, ask yourself: would Van Gogh have wanted to hear the opinion of someone who's never created a master….
0
1K
0