Network Bio partners with Duke University to develop AI models of human disease in USA

USA—Network Bio has partnered with Duke University to develop artificial intelligence (AI) models of human disease using biological samples and patient records, aiming to improve disease detection, drug discovery and treatment selection.

Under the research agreement, the US biotechnology company will collaborate with Duke researchers to conduct molecular profiling of tissue and blood samples stored in the university’s biobanks.

The findings will be combined with de-identified patient information and treatment outcomes, subject to approved research protocols and applicable data privacy regulations.

The integrated datasets will support the development of AI models capable of identifying diseases at earlier stages, uncovering potential therapeutic targets and matching patients with suitable treatments.

Network Bio will also return molecular profiling results to Duke’s biobanks to support further academic research.

Building AI models from human biological data

The collaboration seeks to address a major limitation in biological AI research: the limited availability of datasets that capture how diseases develop and respond to treatment in humans over time.

Many existing models rely on animal studies, laboratory cell lines or publicly available datasets that may not adequately reflect disease progression and long-term clinical outcomes in patients.

By combining physical biospecimens with molecular information and longitudinal medical records, Network Bio and Duke aim to build models based on real-world human biology.

The agreement also establishes arrangements covering data licensing, AI model development and the commercial and scientific use of research outputs.

This framework provides a basis for translating academic biobank resources into datasets that could support pharmaceutical research, biomarker development and diagnostic innovation.

Network Bio co-founder and Chief Business Officer Raphael Potter said the combination of biobanked tissue and blood samples with years of clinical follow-up would help researchers understand how diseases develop and respond to treatment.

For pharmaceutical and diagnostics partners, he explained, the approach could enable researchers to evaluate therapeutic targets, biomarkers and patient selection strategies against human clinical outcomes before advancing programmes into clinical development.

Expanding academic biobank partnerships

Network Bio is developing a patient world model designed to map disease progression and treatment responses across human populations by integrating biological samples with medical histories.

Co-founder and Chief Executive Officer Asad Ali Ahmad said academic medical centres hold extensive collections of molecular data linked to patient outcomes, creating opportunities to improve disease modelling and biomedical research.

He also highlighted the company’s plans to expand its network of academic medical centre biobank partners, which already includes the University of Pennsylvania, Mass General Brigham and the University of Colorado Anschutz.

Network Bio focuses on precision diagnostics, biomarker discovery and drug development by connecting multimodal patient data with machine learning (ML) systems.

Sign up HERE to receive our email newsletters with the latest news and insights from Africa and around the world, and follow us on LinkedIn for updates.