𝗖𝗼𝗵𝗼𝗿𝘁 𝘀𝘁𝘂𝗱𝘆 𝗼𝗳 𝘁𝗵𝗲 𝘄𝗲𝗲𝗸 - The article “𝘋𝘪𝘴𝘦𝘢𝘴𝘦 𝘥𝘪𝘢𝘨𝘯𝘰𝘴𝘵𝘪𝘤𝘴 𝘶𝘴𝘪𝘯𝘨 𝘮𝘢𝘤𝘩𝘪𝘯𝘦 𝘭𝘦𝘢𝘳𝘯𝘪𝘯𝘨 𝘰𝘧 𝘉 𝘤𝘦𝘭𝘭 𝘢𝘯𝘥 𝘛 𝘤𝘦𝘭𝘭 𝘳𝘦𝘤𝘦𝘱𝘵𝘰𝘳 𝘴𝘦𝘲𝘶𝘦𝘯𝘤𝘦𝘴” published in @ScienceMagazine from corresponding authors Anshul Kundaje, @ScottBoydLab and first authors @zazius and Erin Craig developed an integrative model, Mal-ID (Machine Learning for Immunological Diagnosis), to address a gap in current diagnostic medical practices, which do not yet utilize the adaptive immune system’s B cell receptor (BCR) and T cell receptor (TCR) sequences in diagnosis. Mal-ID integrates three models: analysis of overall repertoire composition, convergent clustering of antigen-specific sequences by edit distance and immune receptor sequence features extracted from a large protein language model. Mal-ID outperformed similar classification approaches with AUROCs up to 0.98, representing a major advancement in diagnostic accuracy when applied to a study cohort of 542 samples. This high level of performance was supportive of the model’s analysis of both BCR and TCR repertoire data, which provided more accurate classification than either receptor type alone, potentially reflecting variation in the roles of B cell and T cell responses in different diseases. A diagram of a diagram of a computer model 𝗞𝗲𝘆 𝗙𝗶𝗻𝗱𝗶𝗻𝗴𝘀: 𝗛𝗶𝗴𝗵 𝗗𝗶𝗮𝗴𝗻𝗼𝘀𝘁𝗶𝗰 𝗔𝗰𝗰𝘂𝗿𝗮𝗰𝘆: Mal-ID uses BCR and TCR repertoire data to effectively distinguish six distinct immune states—COVID-19, HIV, lupus, type 1 diabetes, recent influenza vaccination responses, and healthy controls—in a study involving 542 individuals. 𝗨𝘀𝗲 𝗼𝗳 𝗟𝗮𝗿𝗴𝗲-𝗦𝗰𝗮𝗹𝗲 𝗣𝗿𝗼𝘁𝗲𝗶𝗻 𝗟𝗮𝗻𝗴𝘂𝗮𝗴𝗲 𝗠𝗼𝗱𝗲𝗹𝘀 𝗳𝗼𝗿 𝗙𝗲𝗮𝘁𝘂𝗿𝗲 𝗘𝘅𝘁𝗿𝗮𝗰𝘁𝗶𝗼𝗻: Mal-ID leveraged advanced protein language models to extract meaningful immune receptor sequence features for improved diagnostic classification. 𝗧𝗮𝗸𝗲 𝗵𝗼𝗺𝗲: By applying protein language models to BCR and TCR sequence data, Mal-ID demonstrates that AI can be used to realize powerful repertoire-based diagnostics. science.org/doi/epdf/10.1126…

🎉 Congratulations to @KateGrabowski , PhD (Associate Professor of Pathology, Johns Hopkins School of Medicine Faculty) for winning the Apple Vision Pro raffle at our Open House! 🥳 Thank you to everyone who joined us to explore our innovative MIPSA technology and antibody reactome profiling. Follow us and stay tuned for more exciting updates from Infinity Bio! ♾️ #RaffleWinner #Innovation #Biotech #AntibodyProfiling #Antibody #BaltimoreEvents #Baltimore #BiotechStartups

Last week, we hosted students and faculty from Coppin Academy High School. It was inspiring to see Baltimore's best and brightest experience hands-on STEM experiments and immunology learnings- the future looks bright! 💫 #STEM #Baltimore #Immunology #Reactome

Absolutely agree with @EricTopol ! 🙌 Characterizing the "immunome"—from lifetime viral exposures to autoantibody reactivities—provides critical insights into disease mechanisms and biomarker development. We are the leader in comprehensive antibody reactome profiling, and are actively supporting academic and industry translational discovery research 🤝 with our RUO assays: VirSIGHT (all human viruses) and HuSIGHT (all human protein isoforms). While direct-to-consumer applications are certainly on the horizon, today we are helping researchers lay the groundwork for better understanding of antibody reactomes. Check out our RUO libraries here: infinitybio.com/infinity-pro…. 👏♾️ #academia #Antibody #AntibodyReactomeProfiling #AntibodyResearch #BiotechInnovations #HumanStudies #Immunotherapy #BiotechStartups #Biotech