Draft — SPartners Incubator

MOTIVATION LETTER

A drug resistance mutation emerges in a cancer patient. Standard tools need a crystal structure to predict it. That structure does not exist for 82 percent of clinically relevant mutations. The venture predicts resistance 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. This beats the published SOTA of mCSM-lig at approximately 0.70. On SKEMPI 2.0, the current AUROC is 0.634, and the roadmap targets an improvement to 0.70 or higher by fine-tuning on 3,000 SKEMPI mutations.

SPartners Incubator is embedded within Servier's R&D center in Paris-Saclay. Servier is a named partner in the venture's roadmap, with a planned pilot collaboration. The incubator's focus on biotech startups addressing unmet medical needs in oncology and immunology directly matches the venture's target indications. The Paris-Saclay ecosystem includes I2BC, Institut Pasteur, and Sanofi's Gentilly site, all of which are accessible from the incubator's location. This geographic and institutional alignment makes SPartners the ideal launchpad to accelerate the Servier pilot and validate the platform in Servier's labs.

My background combines pharmacy and machine learning engineering. I hold a pharmacy degree and have worked as a pharmacist before transitioning into ML engineering. This dual expertise allows me to understand both the biological mechanism of resistance and the computational approach to predicting it. The venture is pre-seed, proof-of-concept validated, and not yet incorporated. SPartners provides lab access, mentorship, and a direct pathway to collaboration with Servier's R&D teams. The incubator's preference for early-stage ventures that can benefit from lab access and mentorship aligns with the venture's current stage.

The application is complete and submitted before December 1st for Golden Ticket consideration. The venture addresses an unmet medical need in drug resistance prediction, uses an innovative AI-first approach that does not require crystal structures, and has a clear plan to collaborate with Servier. SPartners is the right programme to turn this proof-of-concept into a validated platform with a pharmaceutical partner.

SHORT ESSAY: UNMET MEDICAL NEED

Drug resistance is the primary cause of treatment failure in oncology and infectious disease. In oncology, 30 to 50 percent of patients with solid tumors develop resistance to targeted therapies within 12 months. In antimicrobial resistance, 1.27 million deaths were directly attributable to resistant infections in 2019. Predicting which mutations will cause resistance before they emerge in the clinic would allow drug developers to design compounds that evade resistance, select combination therapies proactively, and prioritize compounds with lower resistance liability.

Current prediction tools require a crystal structure of the protein-drug complex. For 82 percent of clinically relevant mutations, no such structure exists. The venture solves this problem by predicting resistance from protein sequence alone, using ESM-2 protein language model delta-embeddings. This achieves 100 percent mutation coverage compared to approximately 18 percent for structure-limited tools. The model already beats published SOTA on the Platinum benchmark, and the roadmap targets an AUROC of 0.70 or higher on SKEMPI 2.0 through fine-tuning on 3,000 mutations.

The venture's initial focus is oncology, specifically predicting resistance to kinase inhibitors and other targeted therapies. Secondary applications include antiviral resistance and antimicrobial resistance. Servier's R&D priorities in oncology and immunology make them the ideal first pharmaceutical partner. The planned pilot with Servier will validate the platform on Servier's proprietary compounds, creating a direct path to commercial application.

SHORT ESSAY: INNOVATION AND FEASIBILITY

The innovation is the combination of ESM-2 protein language model delta-embeddings with ECFP4 drug fingerprints in a Random Forest classifier. This approach does not require crystal structures, which is the key limitation of existing tools. The model learns the relationship between sequence-level protein features and drug chemical features to predict whether a specific mutation will cause resistance to a specific drug.

Feasibility is demonstrated by the AUROC of 0.804 on the Platinum benchmark, which includes 553 mutations across multiple protein targets. The protein-grouped cross-validation ensures the model generalizes to unseen proteins, not just unseen mutations of known proteins. The lower performance on SKEMPI 2.0 (AUROC 0.634) reflects the greater diversity of that dataset and identifies the specific improvement target for the next development phase.

The roadmap is concrete. Fine-tune ESM-2 on 3,000 SKEMPI mutations to achieve AUROC 0.70 or higher. Validate on an independent set of clinically reported resistance mutations. Then initiate the Servier pilot, using Servier's proprietary compounds and resistance data. The pilot will generate the first revenue and establish the platform's commercial value proposition.

The technical team is me, a pharmacist-turned-ML-engineer with sole authorship of the venture. The SPartners Incubator provides access to Servier's R&D expertise, lab facilities, and mentorship from BioLabs. The Paris-Saclay ecosystem adds academic collaborators from I2BC and Institut Pasteur. This combination of technical proof, clear roadmap, and institutional support makes the venture feasible and fundable.

SHORT ESSAY: TEAM AND EXECUTION

I am Eniola Olutogun, a pharmacist and machine learning engineer. My pharmacy training gives me domain expertise in drug mechanisms, resistance pathways, and clinical relevance. My ML engineering training gives me the technical skills to build and deploy protein language models, train classifiers, and validate performance on biological benchmarks. I am the sole author of the venture, responsible for all technical development, business strategy, and fundraising.

The venture is pre-seed and not yet incorporated. The proof-of-concept is validated on the Platinum benchmark. The next execution steps are: incorporate the venture in France, fine-tune ESM-2 on SKEMPI 3K mutations, achieve AUROC 0.70 or higher, initiate the Servier pilot, and generate first revenue. The timeline is 12 to 18 months.

SPartners Incubator is critical to execution. The incubator provides lab access, mentorship, and direct connection to Servier's R&D teams. The Paris-Saclay location gives access to academic collaborators at I2BC and Institut Pasteur. The incubator's network of biotech mentors and investors supports the transition from pre-seed to seed funding.

The support pipeline is active. SEMIA and Quest for Health meetings are in progress. WILCO One BioTech is scheduled for October 2026. IncubAlliance and AI House applications are submitted. EIC Accelerator and BPI i-Lab are future targets. SPartners is the immediate priority because of its direct alignment with Servier and the Paris-Saclay ecosystem.

CHECKLIST

- [ ] Complete SPartners Incubator application form on msmeafricaonline.com
- [ ] Upload motivation letter (this document)
- [ ] Upload short essay on unmet medical need
- [ ] Upload short essay on innovation and feasibility
- [ ] Upload short essay on team and execution
- [ ] Upload CV for Eniola Olutogun
- [ ] Upload proof-of-concept validation results (AUROC 0.804 on Platinum benchmark)
- [ ] Upload roadmap document with timeline and milestones
- [ ] Upload letter of support or expression of interest from Servier (if available)
- [ ] Submit before December 1st for Golden Ticket consideration

EDITOR NOTES

- Eligibility risk: SPartners Incubator is listed on msmeafricaonline.com, which appears to focus on African startups. Verify that the programme accepts non-African founders or that Eniola's Nigerian origin qualifies. The URL domain suggests a possible Africa-focused programme, which may require the startup to have African operations or founders.
- Fact verification: Confirm that Servier's R&D center in Paris-Saclay is indeed the location of SPartners Incubator. The profile states SPartners is embedded within Servier's R&D center, but the programme URL does not mention Servier. Verify this connection directly with the programme.
- Gap: The profile does not specify Eniola's nationality or residency status. If the programme requires EU residency or French incorporation, clarify whether Eniola has the right to work and incorporate in France. Insert this detail in the team section if applicable.
- Gap: The profile mentions "SEMIA and Quest for Health meeting in progress" but does not specify the outcome or date. If the meeting has occurred, add the result. If not, note it as pending.
- Gap: The profile does not include a specific dollar amount for the SPartners Incubator programme. The application may require a budget or funding request. Prepare a placeholder budget of 30,000 to 50,000 euros for pre-seed validation costs.

MOTIVATION LETTER

A drug resistance mutation emerges in a cancer patient. Standard tools need a crystal structure to predict it. That structure does not exist for 82 percent of clinically relevant mutations. The venture predicts resistance 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. This beats the published SOTA of mCSM-lig at approximately 0.70. On SKEMPI 2.0, the current AUROC is 0.634, and the roadmap targets an improvement to 0.70 or higher by fine-tuning on 3,000 SKEMPI mutations.

SPartners Incubator is embedded within Servier's R&D center in Paris-Saclay. Servier is a named partner in the venture's roadmap, with a planned pilot collaboration. The incubator's focus on biotech startups addressing unmet medical needs in oncology and immunology directly matches the venture's target indications. The Paris-Saclay ecosystem includes I2BC, Institut Pasteur, and Sanofi's Gentilly site, all of which are accessible from the incubator's location. This geographic and institutional alignment makes SPartners the ideal launchpad to accelerate the Servier pilot and validate the platform in Servier's labs.

My background combines pharmacy and machine learning engineering. I hold a pharmacy degree and have worked as a pharmacist before transitioning into ML engineering. This dual expertise allows me to understand both the biological mechanism of resistance and the computational approach to predicting it. The venture is pre-seed, proof-of-concept validated, and not yet incorporated. SPartners provides lab access, mentorship, and a direct pathway to collaboration with Servier's R&D teams. The incubator's preference for early-stage ventures that can benefit from lab access and mentorship aligns with the venture's current stage.

The application is complete and submitted before December 1st for Golden Ticket consideration. The venture addresses an unmet medical need in drug resistance prediction, uses an innovative AI-first approach that does not require crystal structures, and has a clear plan to collaborate with Servier. SPartners is the right programme to turn this proof-of-concept into a validated platform with a pharmaceutical partner.

SHORT ESSAY: UNMET MEDICAL NEED

Drug resistance is the primary cause of treatment failure in oncology and infectious disease. In oncology, 30 to 50 percent of patients with solid tumors develop resistance to targeted therapies within 12 months. In antimicrobial resistance, 1.27 million deaths were directly attributable to resistant infections in 2019. Predicting which mutations will cause resistance before they emerge in the clinic would allow drug developers to design compounds that evade resistance, select combination therapies proactively, and prioritize compounds with lower resistance liability.

Current prediction tools require a crystal structure of the protein-drug complex. For 82 percent of clinically relevant mutations, no such structure exists. The venture solves this problem by predicting resistance from protein sequence alone, using ESM-2 protein language model delta-embeddings. This achieves 100 percent mutation coverage compared to approximately 18 percent for structure-limited tools. The model already beats published SOTA on the Platinum benchmark, and the roadmap targets an AUROC of 0.70 or higher on SKEMPI 2.0 through fine-tuning on 3,000 mutations.

The venture's initial focus is oncology, specifically predicting resistance to kinase inhibitors and other targeted therapies. Secondary applications include antiviral resistance and antimicrobial resistance. Servier's R&D priorities in oncology and immunology make them the ideal first pharmaceutical partner. The planned pilot with Servier will validate the platform on Servier's proprietary compounds, creating a direct path to commercial application.

SHORT ESSAY: INNOVATION AND FEASIBILITY

The innovation is the combination of ESM-2 protein language model delta-embeddings with ECFP4 drug fingerprints in a Random Forest classifier. This approach does not require crystal structures, which is the key limitation of existing tools. The model learns the relationship between sequence-level protein features and drug chemical features to predict whether a specific mutation will cause resistance to a specific drug.

Feasibility is demonstrated by the AUROC of 0.804 on the Platinum benchmark, which includes 553 mutations across multiple protein targets. The protein-grouped cross-validation ensures the model generalizes to unseen proteins, not just unseen mutations of known proteins. The lower performance on SKEMPI 2.0 (AUROC 0.634) reflects the greater diversity of that dataset and identifies the specific improvement target for the next development phase.

The roadmap is concrete. Fine-tune ESM-2 on 3,000 SKEMPI mutations to achieve AUROC 0.70 or higher. Validate on an independent set of clinically reported resistance mutations. Then initiate the Servier pilot, using Servier's proprietary compounds and resistance data. The pilot will generate the first revenue and establish the platform's commercial value proposition.

The technical team is me, a pharmacist-turned-ML-engineer with sole authorship of the venture. The SPartners Incubator provides access to Servier's R&D expertise, lab facilities, and mentorship from BioLabs. The Paris-Saclay ecosystem adds academic collaborators from I2BC and Institut Pasteur. This combination of technical proof, clear roadmap, and institutional support makes the venture feasible and fundable.

SHORT ESSAY: TEAM AND EXECUTION

I am Eniola Olutogun, a pharmacist and machine learning engineer. My pharmacy training gives me domain expertise in drug mechanisms, resistance pathways, and clinical relevance. My ML engineering training gives me the technical skills to build and deploy protein language models, train classifiers, and validate performance on biological benchmarks. I am the sole author of the venture, responsible for all technical development, business strategy, and fundraising.

The venture is pre-seed and not yet incorporated. The proof-of-concept is validated on the Platinum benchmark. The next execution steps are: incorporate the venture in France, fine-tune ESM-2 on SKEMPI 3K mutations, achieve AUROC 0.70 or higher, initiate the Servier pilot, and generate first revenue. The timeline is 12 to 18 months.

SPartners Incubator is critical to execution. The incubator provides lab access, mentorship, and direct connection to Servier's R&D teams. The Paris-Saclay location gives access to academic collaborators at I2BC and Institut Pasteur. The incubator's network of biotech mentors and investors supports the transition from pre-seed to seed funding.

The support pipeline is active. SEMIA and Quest for Health meetings are in progress. WILCO One BioTech is scheduled for October 2026. IncubAlliance and AI House applications are submitted. EIC Accelerator and BPI i-Lab are future targets. SPartners is the immediate priority because of its direct alignment with Servier and the Paris-Saclay ecosystem.

CHECKLIST

- [ ] Complete SPartners Incubator application form on msmeafricaonline.com
- [ ] Upload motivation letter (this document)
- [ ] Upload short essay on unmet medical need
- [ ] Upload short essay on innovation and feasibility
- [ ] Upload short essay on team and execution
- [ ] Upload CV for Eniola Olutogun
- [ ] Upload proof-of-concept validation results (AUROC 0.804 on Platinum benchmark)
- [ ] Upload roadmap document with timeline and milestones
- [ ] Upload letter of support or expression of interest from Servier (if available)
- [ ] Submit before December 1st for Golden Ticket consideration

EDITOR NOTES

- Eligibility risk: SPartners Incubator is listed on msmeafricaonline.com, which appears to focus on African startups. Verify that the programme accepts non-African founders or that Eniola's Nigerian origin qualifies. The URL domain suggests a possible Africa-focused programme, which may require the startup to have African operations or founders.
- Fact verification: Confirm that Servier's R&D center in Paris-Saclay is indeed the location of SPartners Incubator. The profile states SPartners is embedded within Servier's R&D center, but the programme URL does not mention Servier. Verify this connection directly with the programme.
- Gap: The profile does not specify Eniola's nationality or residency status. If the programme requires EU residency or French incorporation, clarify whether Eniola has the right to work and incorporate in France. Insert this detail in the team section if applicable.
- Gap: The profile mentions "SEMIA and Quest for Health meeting in progress" but does not specify the outcome or date. If the meeting has occurred, add the result. If not, note it as pending.
- Gap: The profile does not include a specific dollar amount for the SPartners Incubator programme. The application may require a budget or funding request. Prepare a placeholder budget of 30,000 to 50,000 euros for pre-seed validation costs.

AI Draft — SPartners Incubator