Because Merck's sub likely spent less than $7 million all in, to end up with over 11 per cent of this not yet US-traded Danish company, it is definitively a. . . gnat.
However, as with many of our most recent mentions. . . it has. . . potential, in the white-hot AI powered immuno-oncology targeted therapies space. So. . . we will show the readership the results of our search. [To my knowledge, Merck itself hasn't mentioned this deal anywhere publicly -- at the health care/life sciences confabs, or elsewhere.] Here's a bit from the Evaxion SEC filings, in its "Description of Business" (but, to be clear -- Merck's aims are broader than just cancer vaccines, of course -- the tech is presumed useful, in identifying therapeutic targets as well):
. . .PIONEER™ is our proprietary AI prediction model for the rapid discovery and design of patient-specific-neoantigens used to derive cancer vaccines. It has been shown that neoantigens, which arise from patient-specific tumor mutations, play a critical role in T-cell mediated antitumor immune response. Neoantigens, being absent in normal tissues, are, we believe, ideal cancer vaccine targets because they distinguish themselves from germline proteins and can be recognized as non-self by the immune system. We believe our AI modules within PIONEER™ enable us to efficiently identify and select those neoantigens that will generate a de novo T-cell response leading to significant antitumor effect in each patient. By combining these neoantigens with a purposefully selected delivery modality believed to further enhance this antitumor effect, we design and deliver our vaccines to patients, effectively training their immune systems to target and kill cancer cells with no or very limited adverse effects on healthy non-cancer cells.
Our proprietary AI modules identifying neoantigens within PIONEER™ have been trained using gradient-boosted decision trees, transformers and a conditional generative adversarial network approach on our internally generated data as well as other data, including, but not limited to, next generation sequencing data from tumor samples, mass spectrometry immunopeptidomics, peptide-MHC-binding affinity data, T-cell immunogenicity data and peptide-MHC-binding stability data. We have demonstrated that development and iterative training of our AI modules improves its predictive power in identifying and selecting therapeutic neoantigens. . . .
Got all that? Cool. But the company proposes patient-by-patient DNA sequencing to create personalized medicines. And as we've long said, this it a very high-price, high-stakes approach.
So -- as many of these might. . . this might end up going. . . nowhere (i.e., flaming out) -- but at these prices, it is a bet on a "field horse" -- at Churchill Downs, on the Friday before Derby day. If it comes in -- everyone gets a lobster dinner on Saturday night! Heh! Now you know. Onward.
नमस्ते
No comments:
Post a Comment