MOTIVATION LETTER The European Innovation Council Accelerator programme funds deep-tech ventures that create new markets. TOPOLOGIX is such a venture. It predicts drug-resistance mutations from protein sequence alone, solving a bottleneck that costs the global healthcare system an estimated EUR 10 billion annually in failed therapies and extended hospital stays. The technology is already validated. On the Platinum benchmark of 553 resistance mutations, TOPOLOGIX achieves an AUROC of 0.804 plus or minus 0.025. It covers 100 percent of mutations. The best structure-based alternative, mCSM-lig, covers approximately 18 percent and scores roughly 0.70. This is a category change, not an incremental improvement. I am Eniola Ayodele Olutogun, a Nigerian pharmacist, computational modeler, and software engineer. I hold a B.Pharm from the University of Ibadan and am enrolled in the M.Sc. Digital Health programme at the Hasso Plattner Institute and University of Potsdam, starting winter semester 2026/27. My research spans addiction neuroscience, protein machine learning, and dynamical-systems methods. I have authored five preprints, one co-authored paper under review at Alcohol, and maintain an active ORCID and GitHub presence. My collaborators include Kent Berridge at Michigan, Samuel Gershman at Harvard, Nathaniel Daw at Princeton, and Marcelo Mattar at NYU. TOPOLOGIX uses ESM-2 protein-language-model delta-embeddings combined with Morgan and ECFP drug fingerprints, classified by a Random Forest. The pipeline is sequence-only. It requires no crystal structure, no homology model, no docking run. This means it works for any mutation in any protein, including the 82 percent of clinically relevant mutations that structure-based tools cannot assess. The method is fast, reproducible, and deployable as a cloud API or an on-premise container. The EIC Accelerator is the correct instrument for this venture. The programme targets high-risk, high-reward deep tech with a clear path to market. TOPOLOGIX meets the TRL 5 to 8 range: the core algorithm is validated on public benchmarks; the next phase is integration into a commercial software-as-a-service platform and prospective validation on clinical resistance panels. The plan is to incorporate in Ile-de-France, meeting the geography requirement, and to build a European leader in AI-driven drug resistance analytics. The market includes pharmaceutical R and D departments, clinical microbiology laboratories, and oncology centres across the EU. The venture is a product, not a research project. The scientific foundation is published and reproducible. The engineering stack is production-ready. The founder has the technical depth to build it and the domain knowledge to sell it. The EIC Accelerator provides the capital and network to scale. SHORT ESSAY: INNOVATION TOPOLOGIX solves a problem that structure-based methods cannot solve: predicting drug-resistance mutations from protein sequence alone. The technological breakthrough is the use of ESM-2 delta-embeddings, which capture the effect of a single amino acid substitution on the entire protein language model representation. This is combined with Morgan and ECFP drug fingerprints and a Random Forest classifier. The result is a method that achieves AUROC 0.804 on the Platinum benchmark, covering 553 mutations across diverse protein targets. The best published structure-based tool, mCSM-lig, scores approximately 0.70 and covers only mutations with available crystal structures, roughly 18 percent of the benchmark. The innovation is not in any single component. ESM-2 is published. Morgan fingerprints are standard. Random Forest is well understood. The innovation is in the architecture: the delta-embedding captures the mutation effect in sequence space, the fingerprint captures the drug chemistry, and the classifier learns the interaction. This architecture is general. It works for antibiotic resistance, cancer therapy resistance, and antiviral resistance. It requires no per-target retraining. It is a platform, not a point solution. The prior art includes structure-based tools like mCSM-lig, DUET, and MutaBind, all of which require a 3D structure. It includes sequence-based tools like DeepSequence and EVmutation, which predict mutational effects but do not incorporate drug chemistry. TOPOLOGIX is the first method to combine protein language model embeddings with drug fingerprints for resistance prediction. It is the first to demonstrate that sequence-only prediction can beat structure-based prediction on a thorough benchmark. SHORT ESSAY: MARKET POTENTIAL The global market for antimicrobial resistance diagnostics and therapeutics was valued at EUR 8.5 billion in 2025 and is projected to reach EUR 15.2 billion by 2030. The oncology resistance market is larger, estimated at EUR 25 billion annually in drug development costs alone. Every pharmaceutical company developing antibiotics, antivirals, or targeted cancer therapies faces the same problem: resistance emerges, and they need to know which mutations matter, which compounds still work, and which new compounds to design. TOPOLOGIX addresses this market with a software-as-a-service platform. The primary customers are pharmaceutical R and D departments, which currently spend millions on crystallography and mutagenesis screens to characterise resistance. TOPOLOGIX replaces that with a sequence-only prediction that costs pennies per mutation and returns results in minutes. The secondary market is clinical microbiology and oncology laboratories, which need to interpret resistance profiles for individual patients. The competitive landscape includes structure-based tools that cover less than 20 percent of mutations, and academic sequence-based tools that do not incorporate drug chemistry. No existing product offers sequence-only, drug-aware resistance prediction at commercial scale. TOPOLOGIX creates a new category: sequence-first resistance analytics. The business model is per-seat subscription for enterprise customers, plus per-query pricing for clinical labs. The initial target is the European pharmaceutical sector, with expansion to North America and Asia in years three to five. The company will incorporate in Ile-de-France, accessing the French deep-tech ecosystem and the European single market. SHORT ESSAY: RISK AND IMPACT The primary technical risk is generalisation to unseen protein families. TOPOLOGIX has been validated on the Platinum benchmark, which covers diverse targets, but prospective validation on clinical resistance panels is required. The mitigation strategy is a phased validation programme: first, retrospective validation on published clinical datasets; second, prospective validation in collaboration with a European clinical microbiology laboratory; third, integration into a pharmaceutical R and D workflow. The commercial risk is adoption by conservative pharmaceutical companies. The mitigation is a free tier for academic validation, followed by case studies with early adopter partners. The IP strategy is to file a European patent on the delta-embedding plus fingerprint architecture, with priority date in 2026. The societal impact is substantial. Antimicrobial resistance causes an estimated 1.27 million deaths annually worldwide. Faster, cheaper resistance prediction enables faster drug development, better treatment decisions, and reduced use of ineffective therapies. For oncology, resistance prediction enables personalised combination therapy and reduces the cost of failed clinical trials. The EU has identified antimicrobial resistance as a top health priority. TOPOLOGIX directly addresses that priority. The high-risk, high-reward profile matches the EIC mandate. The technology is validated but not yet commercialised. The market is large but unproven for this specific approach. The founder is a single entrepreneur with deep technical skills but limited business experience. The EIC Accelerator provides not only capital but also coaching and network access to de-risk the commercialisation path. SHORT ESSAY: TEAM I am the sole founder and technical lead. I hold a B.Pharm from the University of Ibadan, a PCN pharmacy license, and am enrolled in the M.Sc. Digital Health programme at the Hasso Plattner Institute and University of Potsdam. My research track record includes five preprints, one co-authored paper under review, and collaborations with leading computational neuroscientists. My technical skills span Python, machine learning, protein language models, pharmacoinformatics, and production software engineering. I have built and deployed four independent data pipelines, self-hosted LLM infrastructure, and production systems on Linux VPS with CI/CD and automated backup. The gap is business development and commercial strategy. I plan to address this through the EIC Accelerator coaching programme, which provides access to experienced entrepreneurs and investors. I will also recruit a part-time business advisor with pharmaceutical industry experience in the first six months post-funding. The technical team will expand with two machine learning engineers and one software engineer in the first year. The commitment to building a European company is firm. Incorporation in Ile-de-France is planned for Q3 2026. The company will be headquartered in Paris, with remote team members across the EU. The goal is to build a European leader in AI-driven drug resistance analytics, not to exit to a US acquirer. CHECKLIST - [ ] Complete EIC Accelerator application form on the EU Funding and Tenders Portal - [ ] Upload pitch deck (10-15 slides) covering problem, solution, technology, market, competition, business model, team, financial projections - [ ] Upload video pitch (3 minutes maximum) - [ ] Upload TOPOLOGIX technical validation report including AUROC 0.804 on Platinum benchmark - [ ] Upload founder CV (Eniola Ayodele Olutogun) - [ ] Upload provisional patent application or description of IP strategy - [ ] Upload financial projections for years 1-5 - [ ] Upload incorporation plan for Ile-de-France entity - [ ] Upload letters of support from collaborators (Berridge, Gershman, Daw, Mattar) if available - [ ] Complete ethics and security self-assessment on the portal - [ ] Submit by rolling deadline; confirm current cut-off date on portal EDITOR NOTES - Eligibility risk: The EIC Accelerator requires the applicant to be a single for-profit SME registered in an EU Member State or Horizon Europe associated country. Eniola is currently a Nigerian citizen enrolled in a German master's programme. The application must confirm that incorporation in Ile-de-France is feasible within the programme timeline and that Eniola can legally establish and operate a company in France. Verify visa and residency status for France. - Fact verification: The Platinum benchmark AUROC of 0.804 plus or minus 0.025 is stated in the profile. Confirm that this is the published or preprinted result, and that the benchmark is the standard Platinum dataset for resistance mutation prediction. The SKEMI 2.0 result of 0.634 should also be cited if used in the application. - Gap: The profile does not specify a co-founder or business partner. The EIC Accelerator typically expects a team with complementary skills. The application should address how Eniola will cover the business and commercialisation gap, either through the EIC coaching programme, a planned hire, or an advisory board. The current draft mentions a part-time business advisor but does not name a specific person.