Jack Michuda retweeted
I must be among an extremely small group of people (n=1?) that have both 1) trained a frontier LLM and 2) designed and synthesized custom viruses in a lab with my own two hands.
And I think that the takes on AI killing us all by creating dangerous viruses is total bogus.
Illustrates nicely the conflation of prognostic and predictive biomarkers. “Predicting overall survival accurately” = prognostic biomarker.
This is not evidence that you can predict counter factuals about treatment outcomes.
With TARIO-2 pretrained patient embeddings, we can predict overall survival accurately from historical pretreatment biopsy H&E alone.
This means that from a simple pathology image, the model can give a probability of a patient benefiting to a therapeutic.
[ C-index reaches a p-value of 0.04 by a 2-sided test relative to null distribution. ]
Jack Michuda retweeted
In one year, I genetically engineered the boring Arabidopsis into a beautiful ornamental flower 🧬🌹
Here is how and why I gave this model organism a visual upgrade 🧵(1/7)
Jack Michuda retweeted
RGB -> depth videos are working in Spatial Media Converter! Next we'll add depth -> stereo -> MV-HEVC so you can view your videos as Spatial Videos on the Apple Vision Pro.
Jack Michuda retweeted
Tweetorial time! We @RecursionPharma mapped consequences of #CRISPR screening of >17K human genes, found a systematic bias confounding all CRISPR screens, traced its molecular cause, and propose a debiasing algorithm.
Jack Michuda retweeted
We’ve released an update to our preprint with @anshulkundaje,
@ScottBoydLab, @robtibshirani, and many others:
“Disease diagnostics using machine learning of immune receptors”
#tweetorial 🧵 (1/9)
New work with @anshulkundaje, @ScottBoydLab, @robtibshirani, and many others:
“Disease diagnostics using machine learning of immune receptors”
#tweetorial 🧵 (1/11)
Jack Michuda retweeted
And more than anything else, what Lin conveyed was that he was a guy listening to music, sharing it with you, and any gimmick or artifice would only spoil the atmosphere of familiarity, relegating it to hokum. Requiescat Lin Bremer, you were famous to us here inside the radius.
Jack Michuda retweeted
Happy to announce the final release of seaborn 0.12.0, a major update with new features that I'm really excited about.
Check out the highlights: medium.com/@michaelwaskom/an…
Read the full release notes: seaborn.pydata.org/whatsnew/…
pip install seaborn==0.12.0
I hope you find it useful!
Jack Michuda retweeted
Will there ever be good digital infrastructure in biotech? Alternate title: "Part of the reason it's hard to hire developers in bio is because they don't have the tools to succeed that they do elsewhere - will that ever change?" statnews.com/2021/11/19/data…
Jack Michuda retweeted
Interested in building an R infrastructure at a fast paced biotech company? Join us and help our computational biologists drive impactful insights for cancer research! 👩💻🧬🧑💻 #rstats #rstudio @RStatsJobsBot @RLadiesColumbus @RLadiesBoston @RLadiesSF tempus.com/about-us/careers/…
Jack Michuda retweeted
Watching the growth of the #DeSci movement has me inspired. Tried channeling some of that in this piece. DMs are open if you'd like to discuss. Huge thanks to @CmanMD @Niklas_TR for their contributions. Happy reading and please share! :) kdtventures.medium.com/how-h…
Jack Michuda retweeted
Preprint from @TempusLabs on predicting homologous recombination repair defect (a marker of responsiveness to platinum & PARP inh drugs) in archival cancer samples, using either targeted DNA sequencing (HRD-DNA), or whole transcriptome (HRD-RNA) analysis.
medrxiv.org/content/10.1101/…
Jack Michuda retweeted
Validation of Genomic and Transcriptomic Models of Homologous Recombination Deficiency in a Real-World Pan-Cancer Cohort medrxiv.org/cgi/content/shor… #medRxiv
Jack Michuda retweeted
You can now order our Tumor Origin (TO) assay along with xT sequencing tests to predict the primary cancer subtype for tumors of unknown/uncertain origin, which together may inform diagnosis & treatment decisions, potentially improving patient outcomes: tempus.co/3p3HaIw.
Jack Michuda retweeted
New paper in @NatureBiotech from @TempusLabs highlights the Tempus xT panel.
Results show extensive molecular profiling + clinical data successfully identified personalized therapies & clinical trials for 76.8% of tested #cancer patients. @jackm12345631
nature.com/articles/s41587-0…
Jack Michuda retweeted
Jackson Michuda ‘17 is hiring a Computational Biologist RNA at @TempusLabs in Chicago, IL. PhD in a quantitative discipline or molecular biology. ow.ly/z71L50mE7pu #DukeAlumniJobs
