.jpeg)
Bio-Symbology · Foundational Research
Detect the transition,
not only the tumour.
Pythology is developing a bio-symbolic research architecture for modelling the biological transitions that may precede invasive cancer and connecting those transitions to evidence-ranked therapeutic mechanisms.
Research hypothesis
Cancer develops through biological transitions.
The central question is whether computational systems can identify combinations of molecular, cellular, spatial and temporal change that indicate movement toward invasive disease before conventional endpoints are reached.
These transitions may involve genomic instability, epigenetic drift, clonal expansion, inflammatory signalling, immune escape, metabolic rewiring and remodelling of the surrounding tissue environment.
Four research pillars
From precursor state to evidence-ranked hypothesis.
The public programme describes the architecture and research questions without publishing code, model weights, thresholds, compound rankings or unpublished candidate identities.
Precursor-State Intelligence
Model recognised premalignant, clonal, dysplastic, metaplastic and high-risk biological states across multiple disease areas.
Progression-Risk Modelling
Investigate which longitudinal, molecular and spatial features may distinguish stable abnormalities from dangerous trajectories.
Biological Mechanism Mapping
Connect observed states to signalling, DNA repair, immunity, epigenetics, inflammation, metabolism and tissue context.
Therapeutic Intelligence
Map precursor state → biomarker → pathway → target → modality → evidence without equating computational relevance with clinical suitability.
What bio-symbolic means
Learned patterns with explicit biological reasoning.
The architecture combines pattern-learning systems with structured representations of pathways, targets, regulatory context, evidence and contradiction.
Neural pattern detection
Learn representations from molecular graphs, omics profiles, pathology, imaging and longitudinal measurements.
Symbolic pathway reasoning
Represent explicit relationships between biomarkers, pathways, mechanisms, targets, modalities and disease context.
Causal consistency checks
Test whether candidate explanations are compatible with known biology, temporal order and competing mechanisms.
Evidence-ranked hypotheses
Return traceable research hypotheses with provenance, uncertainty, contradictions and validation status.
Initial research programmes
Where precursor biology is already meaningful.
The programme begins with disease areas in which precursor or high-risk states are clinically and biologically recognised, while treating progression prediction as a separate problem.
Precursor Atlas
Colorectal adenomas, serrated lesions, inflammatory dysplasia, Barrett’s-associated change and gastric metaplasia.
Pre-invasive Biology
Atypical hyperplasia, LCIS, DCIS and molecular features associated with heterogeneous progression risk.
Clonal Evolution
CHIP, CCUS, MDS, MGUS and smouldering myeloma studied through longitudinal clone and pathway behaviour.
Cystic and Intraepithelial States
IPMN, mucinous cystic neoplasms and pancreatic intraepithelial progression where risk stratification is difficult.
Virus-Associated Carcinogenesis
Persistent infection, host response, immune escape and cellular transformation in selected oncogenic viruses.
Field Cancerisation
Broader molecular and cellular changes surrounding visible lesions or occurring before a discrete lesion is obvious.
Regulatory-linked molecular knowledge
More than a SMILES file.
The compound layer is designed as a machine-readable knowledge base connecting structure, target, mechanism, pathway, indication, modality, regulatory status and evidence source.
Chemical representations
Canonical and isomeric SMILES, InChI, molecular identifiers, stereochemistry and physicochemical metadata where meaningful.
Separate modality model
Antibodies, cell therapies, gene therapies, vaccines, radiopharmaceuticals and ADCs are represented through modality-specific fields rather than forced into a small-molecule schema.
Public findings policy
Show the work without exposing the crown jewels.
Public mission summaries can report scope, convergence, evidence distribution and validation stage without publishing proprietary candidate rankings or exact relationships.
Programme-level outputs
Disease category, pathways examined, sources integrated, target classes, model agreement and evidence grades.
Protected research detail
Compound identities, target–compound rankings, binding values, feature weights, thresholds and unpublished graph paths.
Evidence status
Established biology, regulatory evidence, clinical evidence, translational evidence, preclinical evidence, computational hypothesis or conflicting evidence.
Scientific restraint
Detection is not destiny.
Not every precursor state progresses. Some remain stable, some regress and others carry highly variable risk. A useful system must distinguish biological danger from mere abnormality while accounting for overdiagnosis, false positives and heterogeneity.
Research foundations
Public scientific and regulatory sources.
External sources provide data and definitions only. Their use does not imply partnership, endorsement or validation of Pythology research.
Cancer molecular characterisation
Genomic, epigenomic, transcriptomic and proteomic data across 33 cancer types.
Open source →Classification of Tumours
Recognised pathology and disease-classification foundations.
Open source →Approval records and labels
Regulatory status, indications, labels and approval history.
Open source →Approval developments
Oncology and haematology approval notifications, checked against Drugs@FDA.
Open source →Biological products
Licensed biological products, reference products and biosimilars.
Open source →Molecular structures
Programmatic chemical identifiers, structures and properties through PUG-REST.
Open source →Research collaboration
Independent validation is the value inflection.
Pythology is interested in scientific discussion with oncology researchers, pathology groups, computational biologists, translational laboratories and institutions able to evaluate well-defined hypotheses.
Pythology Bio-Symbology is an independent computational research programme. It is not a medical device, diagnostic service or clinical decision-support system. Research outputs do not constitute medical advice or treatment recommendations. All hypotheses require independent laboratory, clinical and regulatory validation.