The Data Delusion, and a Path to Computable Biology
KGC 2026
•
33m
https://drive.google.com/file/d/1sbw8uXzA-hR1hvBPzVAeUD-SK7gpAgwi/view?usp=share_link
For decades, the Life Sciences industry has operated under a "Data Delusion": the belief that if we simply aggregate enough "clean" data and apply it at scale, biology will become computable. Yet, despite massive investments in lab automation and data lake curation, drug discovery remains a lottery governed by Eroom’s Law. This talk argues that our current datasets are not only infinitesimally sparse and trivially small, but that the act of "cleaning" data often erases the very signals required for breakthrough insights. To break the cycle, we must move beyond statistical pattern matching and toward a new Resolution Paradigm.
The path forward lies in Large Quantitative Models (LQMs): Grounded world models that incorporate the fundamental laws of physics and information theory. By shifting from an empirical "brute force" mindset to one of first-principles resolution, we can transform drug discovery. We will build on the successful example of protein folding to examine specific next steps (e.g. quantum enzymology, allosteric signaling), and we will explore how Knowledge Graphs serve as the essential connective infrastructure for producing Manifest Evidence and realizing the dream of computable biology.
Up Next in KGC 2026
-
IDMP Ontology in Action
https://www.dropbox.com/scl/fo/p5oi87h9800o35bhyua80/AHipulxYW0IQJrinOTB-Qy0?rlkey=x5fvfx8gsusubrqkx55h0obkp&e=5&st=nxa2ste2&dl=0
We intend to give the presentation on site. This is a shared proposal together with my colleague Dr. Fabian Muttach from Regulatory Affairs. ( https://www.linkedin.co...
-
Human Expertise, Agent Intelligence
The expertise your teams carry — domain rules, entity relationships, business logic — is exactly what AI agents need to make reliable decisions. A knowledge graph makes that expertise queryable: not document search, but a structured, traceable model that agents reason over directly. In this hands...
-
From Knowledge Graphs to Digital Twin...
https://drive.google.com/file/d/1db06jIurJyQDtDwgCOfCLAuy3E_eB3gV/view?usp=share_link
The life sciences community is adopting emerging modeling approaches—LinkML, SHACL-centric design, and updated standards like RDF 1.2—while simultaneously pushing toward ambitious paradigms such as digital twin...
