AI that interrogates evidence—
not a black-box predictor.
PDAO™ AI is Oncotelic’s evidence-interrogation platform. Instead of training a bespoke model to imitate your dataset, we structure, embed, cluster, and query large bodies of biomedical knowledge so the dataset can speak—yielding reproducible, auditable, testable hypotheses.
Interrogation layer (no bespoke model training)
We prioritize evidence organization and retrieval over training a custom black-box predictor. The goal: surfaced signals you can validate.
Reproducible and auditable
Semantic retrieval + clustering + query traces provide a clear chain from question → evidence → hypothesis.
Translation-oriented outputs
PDAO™ AI generates hypothesis candidates: pathway dependencies, tumor contexts, biomarker patterns linked to endpoints.
How it works (at a glance)
Ingest literature + structured datasets.
Embed into a semantic space (similarity).
Cluster recurring themes to expose patterns.
Query reproducibly to generate hypotheses.
Output is evidence-linked hypotheses—designed to be testable against biology and outcomes.
TGF-β Corpus: comprehensive, conversant knowledge
Oncotelic has curated a comprehensive TGF-β literature corpus containing 125,000+ PubMed abstracts—representing the totality of scientific knowledge related to TGF-β across oncology, immunology, fibrosis, metabolism, and translational therapeutics.
The corpus is embedded, clustered, and indexed in PDAO™ AI so researchers can interrogate the evidence to generate new hypotheses and new insights.
What researchers can do
Hypothesis mining: identify recurring mechanistic links around endpoints (survival, resistance, immune escape).
Context mapping: discover tumor microenvironment contexts that repeatedly co-occur with TGF-β signals.
Cross-domain synthesis: connect immunology, fibrosis, metabolism, and oncology without being trapped by one training set.
Evidence traceability: maintain a chain from question → retrieved evidence → proposed hypothesis.
PDAO™ AI → TGF-β Corpus → Hypothesis → Translation
Discord access
The TGF-β corpus community is hosted on Discord to support collaborative exploration, hypothesis discussion, and onboarding.
Onboarding guide (researchers)
A lightweight path to get value in 15 minutes once you join Discord.
Join Discord and read the welcome + rules.
Introduce yourself: tumor area, focus (TME, fibrosis, immunology), and what you want to discover.
Start with 3 queries: (i) TGFB2 + survival, (ii) TGF-β + macrophages, (iii) TGF-β + resistance.
Capture outputs: cluster summaries + top abstracts + your testable hypothesis statement.
Share & iterate in the hypotheses channel: what recurs, what contradicts, and what to validate next.
Generate a copyable access request
Use this to route inbound interest internally (research, collaboration, investor diligence).
