Framing Angle (from Research)
For Eniola Olutogun, the strongest frame is: 'the only resistance-prediction tool that works on day one of a drug program, before any crystal structure exists.' The 100% mutation coverage vs 17.6% for structure-limited tools is the killer stat for a Novartis or Roche reviewer who knows that novel-target programs spend years waiting for structural data before they can screen resistance risk — position the venture as solving a bottleneck that is invisible to standard tooling and that directly affects go/no-go decisions in pharma R&D. Frame the ESM-2 delta-embedding approach as TechBio-native computational infrastructure (not a one-off classifier), and anchor credibility to the Platinum benchmark result (AUROC 0.634, protein-grouped CV — methodology-rigorous), the Servier pilot in progress, and Paris-Saclay/Institut Pasteur collaborators. By May 2027, lead with AUROC ≥0.70 post-fine-tuning and pilot data, and pre-position the ask as: 'DayOne's Roche/Novartis network is the exact validation environment we need to prove generalisability across novel oncology and antiviral targets before scaling.'
Full Research →
The application materials are saved to .
Here is what the file contains and a few things worth knowing before you read it:
WHAT IS IN THE FILE
Six complete sections: Venture Summary, Problem and Solution Statement, Technical Overview, Team Profile, Pitch Video Script (3-minute spoken script, timed), and Why DayOne, plus a Checklist and Editor Notes at the end. Each is sized to the format it targets, from the 150-word form field to the 400-word technical document.
THE CORE FRAMING USED THROUGHOUT
The "day one of a drug program" angle is the spine. Every section returns to the same contrast: structure-limited tools cover 17.6% of Platinum benchmark mutations; this system covers 100%, because it needs no crystal structure. The Roche and Novartis hook in the pitch video and in the Why DayOne essay is deliberately specific to their novel-target pipeline cadence, not to pharma in general.
WHAT TO ACT ON BEFORE SUBMITTING IN MAY 2027
The Editor Notes section flags the four most important things to fix:
1. Incorporate a legal entity, SAS or equivalent, before the application opens. DayOne requires a legal counterparty for its cohort agreement.
2. Do not forward-project AUROC 0.70 if the ESM-2 fine-tuning is not done by application date. The technical overview as drafted presents it as a milestone, not a current result, but double-check the language when you submit.
3. Qualify the Servier reference to match whatever is actually true by May 2027. A signed LOI becomes an asset; an informal conversation should be described as "pilot discussions."
4. The source brief listed "Novartis/Novartis network" twice. Confirm DayOne's actual 2027 mentor network from the live website when the application opens, and name the relevant companies precisely.