MOTIVATION LETTER Drug resistance is the single largest cause of treatment failure in oncology, antivirals, and antimicrobial therapy. Current prediction tools require a protein crystal structure, which exists for fewer than 20 percent of clinically relevant mutations. My venture solves this: it predicts drug resistance mutations from protein sequence alone, using ESM-2 protein language model delta-embeddings combined with ECFP4 drug fingerprints and a Random Forest classifier. On the Platinum benchmark of 553 mutations, the model achieves an AUROC of 0.804 plus or minus 0.025 under protein-grouped cross-validation, outperforming the published state-of-the-art mCSM-lig at approximately 0.70. It covers 100 percent of mutations versus roughly 18 percent for structure-limited tools. I am Eniola Olutogun, a pharmacist turned machine-learning engineer. I built this venture as a sole author and validated the proof-of-concept independently. The Women-Led Business Grant 2026 is the right programme at the right moment. The grant provides non-dilutive funding that will directly support three critical milestones: compute credits for fine-tuning ESM-2 on the SKEMPI 3K mutation dataset to reach an AUROC of 0.70 or higher, a pilot deployment with Servier at their Suresnes site, and initial technology procurement for the Servier pilot. These steps convert a validated proof-of-concept into a revenue-generating product. The programme’s focus on empowerment, leadership, and innovation aligns with my trajectory. I transitioned from clinical pharmacy to computational biology without a formal computer science degree, teaching myself protein language models and building a model that beats published benchmarks. The venture addresses an urgent global health need: predicting resistance before it emerges in the clinic, enabling smarter drug design and longer-lasting therapies. The grant’s support for pilot deployments and technology procurement matches my immediate capital requirements. I am targeting incorporation in Ile-de-France, with named partners including Servier, Paris-Saclay, and Sanofi. My support pipeline includes SEMIA, WILCO One BioTech, and applications to IncubAlliance and AI House. The Women-Led Business Grant will accelerate my timeline by funding the compute and pilot costs that bridge proof-of-concept to first revenue. SHORT ESSAY: INNOVATION AND IMPACT The venture’s core innovation is removing the crystal-structure requirement from drug resistance prediction. Existing tools like mCSM-lig depend on a 3D protein structure, which is unavailable for the vast majority of clinically observed mutations. My model uses ESM-2 embeddings to capture sequence-level evolutionary information, combined with ECFP4 drug fingerprints to encode the small-molecule side of the interaction. The Random Forest classifier then predicts whether a given mutation will confer resistance to a given drug. On the Platinum benchmark, the model achieves an AUROC of 0.804 with a standard deviation of 0.025 under protein-grouped cross-validation. This beats the published mCSM-lig AUROC of approximately 0.70. On the harder SKEMPI 2.0 benchmark, the AUROC is 0.634, which sets the baseline for the fine-tuning work this grant will fund. The model covers 100 percent of mutations in the benchmark, compared to roughly 18 percent for structure-dependent tools. The impact spans oncology, antiviral development, and antimicrobial resistance. In oncology, resistance to targeted therapies like kinase inhibitors emerges within months. In antivirals, resistance mutations in HIV and influenza protease active sites reduce drug efficacy. In antimicrobial resistance, predicting resistance in bacterial proteins can guide antibiotic stewardship. The venture’s technology applies to any protein-drug pair, making it a platform rather than a single-target tool. SHORT ESSAY: FUNDING USE AND BUSINESS GROWTH The grant funds will be allocated to three areas. First, compute credits for fine-tuning ESM-2 on the SKEMPI 3K mutation dataset. This requires approximately 2,000 GPU-hours on an A100-equivalent cluster, estimated at 15,000 euros. Second, the Servier pilot deployment, which includes data integration, validation runs, and reporting, estimated at 25,000 euros. Third, technology procurement for the pilot, including software licenses and cloud infrastructure, estimated at 10,000 euros. Total direct costs are 50,000 euros. The remaining funds will support working capital for incorporation in France, legal fees for intellectual property protection, and market research into the EU biopharma landscape. The business model is software-as-a-service with per-target licensing to pharma R&D teams. The first target customer is Servier, with a pilot that will generate the first annual recurring revenue. After the pilot, the venture will expand to Sanofi and other Paris-region biopharma companies. The grant’s non-dilutive structure is critical at the pre-seed stage. It preserves equity for future institutional rounds while funding the technical milestones that de-risk the venture for later investors. The Women-Led Business Grant directly enables the transition from sole-author proof-of-concept to incorporated company with a paying customer. RESEARCH STATEMENT The venture’s research programme has three phases. Phase one, completed, is the proof-of-concept: a Random Forest classifier using ESM-2 delta-embeddings and ECFP4 fingerprints, validated on the Platinum benchmark with an AUROC of 0.804. Phase two, funded by this grant, is fine-tuning the ESM-2 model on the SKEMPI 3K mutation dataset to improve performance on binding affinity changes. The target is an AUROC of 0.70 or higher on SKEMPI 2.0. Phase three is deploying the fine-tuned model in a pilot with Servier, predicting resistance mutations for an undisclosed oncology target. The technical approach uses delta-embeddings, which capture the change in protein representation between wild-type and mutant sequences. This is computationally efficient and does not require multiple sequence alignment or structural modeling. The ECFP4 fingerprints encode drug molecular structure as a binary vector. The Random Forest classifier is interpretable, which is important for regulatory acceptance in pharma. The research is conducted in collaboration with the Institute for Integrative Biology of the Cell at Paris-Saclay and the Institut Pasteur. The Servier pilot will provide real-world validation data and feedback for model refinement. The long-term research goal is to extend the model to multi-drug resistance prediction and to incorporate clinical outcome data from published trials. CHECKLIST - [ ] Motivation letter, 300-500 words, submitted as PDF - [ ] Short essay on innovation and impact, 200-350 words, submitted as PDF - [ ] Short essay on funding use and business growth, 200-350 words, submitted as PDF - [ ] Research statement, 400-600 words, submitted as PDF - [ ] Founder CV for Eniola Olutogun - [ ] Proof of women-led business status (certificate of incorporation or founder declaration) - [ ] Budget table showing allocation of grant funds to compute, pilot, and procurement - [ ] Letters of support from Servier and Paris-Saclay (in progress) - [ ] Submit via programme website at https://opportunitiesforyouth.org/2026/06/21/2025-womens-opportunities-for-girls-women-start-ups-and-ngos-empowerment-leadership-and-innovation-apply-now/ EDITOR NOTES - Eligibility risk: The programme requires the business to be at least 51 percent women-owned or have women in C-suite leadership. The venture is not yet incorporated. Eniola should confirm whether a founder declaration of intent to incorporate as women-led is acceptable, or whether incorporation must be completed before submission. - Fact verification: The SKEMPI 3K dataset size and availability should be confirmed. The profile states 3K mutations, but the public SKEMPI 2.0 dataset contains approximately 3,000 entries. Verify that the fine-tuning dataset is accessible and licensed for commercial use. - Gap to fill: The profile does not specify the exact grant amount requested. Eniola should decide a specific figure between 25,000 and 50,000 euros based on the budget outlined in the funding use essay, and state it clearly in the motivation letter and budget table.