JP Morgan’s annual Healthcare Conference is becoming almost an AI event as doctors and AI scientists collaborate to turn petabytes of clinical data into generative AI models and actionable insights. The progress made since the advent of Chat GPT shows that we are on the precipice of a revolution that will produce cures for many common diseases and save countless lives.
Cerebras Systems (a Cambrian-AI Research client) announced a new baseline model co-developed with partner The Mayo Clinic that can identify the best medical therapy for Rheumatoid Arthritis.
Cerebras and Mayo Clinic apply AI to healthcare to predict drug efficacy
These two institutions have collaborated for more than a year and produced a new foundational model that can predict the efficacy of drugs to relieve Rheumatoid Arthritis (RA). Traditionally, RA sufferers resorted to a trial-and-error approach, with each trial testing a drug for six months. If it doesn’t work, move on to the next drug and repeat until you hopefully find one that works. Meanwhile, your RA is getting worse.
The genomic model was trained by combining publicly available human reference genome data with Mayo’s comprehensive patient exome (protein coding) data. By feeding this data into the Cerebras wafer-scale engine, the companies have now trained the first LLM based on genomic data to predict which drug will most benefit a particular RA patient. The Human Reference Genome is a digital DNA sequence that represents a composite, “idealized” version of the human genome.
Unlike other models trained exclusively on the human reference genome, the Mayo Genomic Foundation model is trained on the Mayo’s Tapestry database that includes 500 unique patient genomes. The combination of the reference and the patient genome creates the primary database for training a genomic foundation model.
Mayo will continue to refine the model with additional genome data to improve accuracy, but it has already achieved 87% accuracy in predicting drugs. Also, the accuracy of predicting the predisposition of cancer and cardiovascular diseases is examined in the chart below.
Where is this going?
These two companies have only been working together for about a year, and are now turning their expertise and computing power to expanding the foundation’s model to include support for other areas of medicine.
The model can be expected to be extended to a variety of tasks, from Pathology to Radiology to genomic research for cancer and other drugs.
Combining the power of Cerebra with data from Mayo is undoubtedly helping to create the next generation of medicine.
FINDINGS: This article expresses the opinions of the author and should not be taken as advice to buy or invest in the companies mentioned. Cambrian-AI Research is fortunate to have many, if not most, semiconductor firms as our clients, including Blaize, BrainChip, Cadence Design, Cerebras, D-Matrix, Eliyan, Esperanto, Flex, GML, Groq, IBM, Intel, Nvidia, Qualcomm Technologies, Si-Five, SiMa.ai, Synopsys, Ventana Microsystems, Tenstorrent and many investment clients. We have no investment positions in any of the companies mentioned in this article and do not plan to initiate any in the near future.
For more information, please visit our website at https://cambrian-AI.com.