is GPT-4v susceptible to optical illusions? hunch is *of course not*...why would a computer vision system exhibit the same "feature/bugs" specifics as the idiosyncratic wetware wiring of the human visual system? yet for some, it weirdly does:

(1/10) Our work on antibody developability is out! We present a cartography of the developability landscapes of the native and human-engineered antibodies. This study was a team effort of Habib Bashour Eva Smorodina Matteo Pariset , and Jahn Zhong e/🧬 biorxiv.org/content/10.110…

Excited to share major progress on next generation of AlphaFold! Together with Isomorphic Labs, we’ve significantly advanced accuracy & expanded coverage beyond proteins to other key biomolecular classes. Continuing to accelerate science at digital speed dpmd.ai/x-alphafold-pr…

(1/7) Curious about what a blood sample can reveal about humoral immunity? Our latest work combined sequencing methods and mass spectrometry to explore BCR and antibody repertoires and assess their synergistic potential. biorxiv.org/content/10.110…

Our study on benchmarking and integrating human B-cell receptor genomic and antibody proteomic profiling is now on bioRxiv biorxiv.org/content/10.110…. Great work by Khang Le Quy Igor Snapkow and Maria Chernigovskaya. All details are below in the 🧵.

I got into ml because the other option was being a lab hazard

Happy to share a new manuscript! "Germinal centers are permissive to subdominant antibody responses" In brief, in simulations, low immunogenic antigens get their B cell response despite the presence of immunodominant antigens. frontiersin.org/articles/10.33… Explanations below!

I'm curious how many early-career scientists (MSc/PhD/PostDoc) in Norway are interested in bio-related entrepreneurship. Do you have a general appeal, do you want to learn more, or maybe you are already developing your idea and launching a startup? UiODoc Universitetet i Oslo @UiO_LifeSci

Lin Yun-Ju and Dimitrij Ovtcharov's skills at the table lit up #ITTFWorlds2024 today 🔥 #Busan2024 #TableTennis #PingPong #Liebherr

New opening for a PhD student in my lab, co-supervised with Fridtjof Lund-Johansen . This position is ERC-funded and focuses on novel experimental technologies for antibody-antigen binding screening 📷 jobbnorge.no/en/available-j…

We’re excited to share our preprint where we show, for the first time, the atomically accurate design of VHH antibodies!

Are you developing new machine learning methods for immune receptor/repertoire data? You can save yourself a lot of time by developing your method inside immuneML. Read all about it in our updated documentation: docs.immuneml.uio.no/latest/develop… (1/3)

Here is my biggest contribution to #science so far. A short series of memes inspired by #AlphaFold3 from Google DeepMind 1/9🧵

Some observations from my very brief #AlphaFold3 (AF3) benchmarking: AF3 is good for globular proteins, soluble proteins, conserved proteins, RNA/DNA-protein complexes, etc. AF3 is bad for IDPs, protein-protein complexes, antibodies, protein-switches, membrane proteins, etc.

Vincent Van Deuren (University of Antwerp) analyzed the synovial fluid T-Cell Receptor (TCR) repertoire across different arthritis groups and T-cell subsets, to characterize the T-cells associated with HLA-B27 related spondyloarthropathies. #AIRRC7 tinyurl.com/airrcmeeting7

Our paper on the linguistics-based formalization of antibody language is out in Nature Computational Science. Work led by Mai Ha Rahmad Akbar and Philippe A. Robert . Summary thread: x.com/PRobertImmodel….

1/8 Can negative dataset choices boost ML performance? We show that selecting negative samples is crucial for enhancing generalization and rule discovery in adaptive immune receptors. Learn more in this thread! 🧬 biorxiv.org/content/10.110…

Exciting news! Our research paper on benchmarking BCR sequencing and antibody mass spectrometry on bulk and single cell levels, previously shared as a preprint, is now officially published. Take a look at the peer-reviewed version here: nature.com/articles/s4154…

This is an interesting idea: turning a cancer's drug resistance against it These are growth curves of tumours in mice, treated with a drug (a tyrosine kinase inhibitor). The tumours initially shrink but they all - slowly but surely - become resistant: This study aimed to convert

Finally biologists can also use numpy (array programming). Handling e.g. DNA and protein sequences with convenience and speed, like physicists and machine learners for decades have worked with numerical data: nature.com/articles/s4159… (1/3)