MallikharjunaSakhamuri/graph-immunology-thesis

This project is a classic PhD thesis repository: high in domain specificity and academic rigor, but low in software defensibility. With 0 stars and forks after 4 months, it lacks any community traction or developer ecosystem. It functions primarily as a 'code dump' for a thesis, making it a reference implementation rather than a deployable tool. The defensibility score is low because the code can be easily cloned and the logic replicated by any researcher in the field; the 'moat' is the author's individual expertise, not the repository itself. While frontier labs like OpenAI or Google DeepMind are unlikely to build this specific niche tool (low frontier risk), specialized AI-biotech companies (e.g., Immunai, Generate:Biomedicines) or established academic tools like NetMHCpan represent significant competition. The displacement horizon is short because research in AI for drug discovery moves rapidly, and specialized hypergraph models are likely to be superseded by more generalizable protein foundation models or better-maintained open-source frameworks within the next 18-24 months. The primary value lies in the niche application of hypergraphs to HLA binding, which is a novel combination of techniques, but without a maintenance roadmap or user base, it remains a fragile academic artifact.