deboramarks: Can AI help predict pancreatic cancer earlier? nature.com/articles/s41591-0… and hms.harvard.edu/news
deboramarks: RT @sara_mostafavi: MLCB 2023, coming back in November 29-30 as a hybrid conference (Seattle !). Check out the fantastic list of invited speakers @elhamazizi @bidumit @alexrives : sites.google.com/cs.washingt…
with great co-organizers @suinleelab @QuonBio @anshulkundaj…
deboramarks: Applicable to other viruses- disease X- to anticipate immune escape - no antibody data needed xitter.cf/nthadanin/status/16505…
We’ve updated our EVEscape paper with more SARS2 retrospective evaluation and potential use-cases! biorxiv.org/content/10.1101/… Collaborators @sarahgurev @NotinPascal @nooryoussef03 @_nathanrollins @sandercbio @yaringal (@OATML_Oxford), @deboramarks, 🧵1/4
deboramarks: RT @AlanNawzadAmin: The preprint of my work with Eli Weinstein and Debbie Marks is up! In it we show that your off-the-shelf kernel for biological sequences is probably not as flexible you thought and build alternative sequence kernels that … arxiv.org/abs/2304.03775
deboramarks: Check out this job at Harvard Medical School: Project Manager, Predictive Modeling linkedin.com/jobs/view/35674…
deboramarks: RT @ml4proteins: Next week on 4/11 @ 4 pm EST, we'll have @NotinPascal talk about hybrid protein language models for fitness prediction!
ml4proteinengineering.com/ap…
deboramarks: Interested in making AI work for Bio and sustainability? Join the community!
discord.gg/RWgxbjZp
deboramarks: @LubanLab So naming goes by genome sequence identity?
deboramarks: @LubanLab Why not SARS-CoV-3? What are the virology rules that I need to learn?
deboramarks: RT @NotinPascal: We now have a dedicated repo for the initiative: github.com/OATML-Markslab/Pr…. You are aware of a relevant assay that we have not yet included? Your favorite model is not there yet? Please see the repo instructions for suggestions of new assays & models
deboramarks: RT @NotinPascal: With 94 assays total, ProteinGym is the largest curated collection of DMS assays to assess the ability of protein models to predict the effects of mutations -- but this is just the beginning, and we will continue to add new …
deboramarks: RT @NotinPascal: A few updates regarding the ProteinGym benchmarks! proteingym.org
deboramarks: @jmschreiber91 Worse than that - many methods cannot discover anything because ‘benchmarks’ are just inadequate for the goal
deboramarks: RT @igisci: Happy #LGBTSTEMDay! Our artist @mpk_draws (she/they) was inspired to promote queer representation by illustrating some contemporary queer STEM researchers – keep reading to find out a little about each person pictured! 🏳️🌈🧵 1/6
… xitter.cf/igisci/status/15936516…
deboramarks: @anshulkundaje Harvard - not officially allowed whatever the funding source. Been a huge problem during Covid …
deboramarks: @kulesatony I like Abbas,Lichtman and Pillai cos it’s v simple
deboramarks: @JohnAllenPaulos Equivalent to the white and black counters problem. Nice.
deboramarks: Looking to buy a stock of swabs for nasal samples - first use case advertised on A is 'Gun Cleaning' - what does that tell us? xitter.cf/deboramarks/status/157…




