04 · 21 proprietary layers

Scientific framework
Explore the 21 analytical layers.
Try it · Ask the layers
Ask the layers a question.
Each layer exists to answer a specific question — and each states what it cannot establish.
Answered by layer 2
SourceGuard
Somatic and germline interpretation
Classification or context of findings, and items that may need confirmation or clinical genetics review.
What it does not establish
A 'possibly germline' flag is not a hereditary diagnosis. Confirmatory germline testing with appropriate consent and counselling is required.
How to read this directory
Names are official. Explanations are layered.
Each layer page then explains the question it addresses, its intended output, what it draws on and — just as important — what it does not establish.
Not every layer runs for every case. Which layers contribute depends on the testing pathway and the information available. Every layer output is presented for the treating oncologist's review; none is a decision.
Output types used on every layer
- Assay-derived finding. Reported from laboratory sequencing data, subject to assay scope and quality.
- Computational inference. Estimated by analysis of available data rather than measured directly.
- Evidence-linked interpretation. Connects a finding to published or curated evidence, which has a source and a date.
- Clinical-context integration. Depends on clinical information supplied by the treating team.
- Twin workflow function. Part of how the Twin is created, reviewed or updated, rather than an analytical result.
- Structured record. A governed record of what was reviewed or decided.
Directory
Search and filter all 21 layers
Showing 21 of 21 layers
Question it addresses
What tumour subpopulations or evolutionary patterns may be represented in the available molecular evidence?
A tumour is rarely one uniform population of cells. It usually contains related subpopulations (subclones) that share some alterations and differ in others. CloneTrace is intended to describe what the available data suggest about that structure: which alterations appear to be carried by most tumour cells, and which may be present in only a fraction of them.
Main output
Potentially relevant genomic findings and a structured summary of the clonality or evolutionary interpretation.
What it does not establish
It does not establish the complete clonal architecture of the tumour, and it does not predict how the tumour will evolve.
Draws on
- Molecular & genomic
- Tumour biology
Output type
- Computational inference
- For treating-oncologist review
Report section
02 Molecular findings
Full layer detailCapability
Understand
Question it addresses
Is a finding interpreted as tumour-acquired, potentially inherited, or unresolved in the context of available samples and workflow?
Not every variant found in tumour DNA was acquired by the tumour. Some are inherited and present in every cell of the body. SourceGuard is intended to flag whether a finding looks tumour-acquired (somatic), potentially inherited (germline) or unresolved, because each needs a different interpretation framework and has different implications.
Main output
Classification or context of findings, and items that may need confirmation or clinical genetics review.
What it does not establish
A 'possibly germline' flag is not a hereditary diagnosis. Confirmatory germline testing with appropriate consent and counselling is required.
Draws on
- Molecular & genomic
- Patient factors
Output type
- Computational inference
- Assay-derived finding
- For treating-oncologist review
Report section
02 Molecular findings
Full layer detailCapability
Understand
Question it addresses
What mutational patterns may be present, and what could they indicate in an appropriate evidence context?
Processes that damage DNA, such as tobacco smoke, ultraviolet light, defective DNA repair or some earlier treatments, tend to leave characteristic patterns of mutations. SigPrint is intended to describe which of these patterns may be present in the tumour's mutations.
Main output
Relevant signature observations and evidence-linked interpretation, where available.
What it does not establish
It does not prove what caused an individual's cancer.
Draws on
- Molecular & genomic
Output type
- Computational inference
- Evidence-linked interpretation
- For treating-oncologist review
Report section
02 Molecular findings
Full layer detailCapability
Understand
Question it addresses
How might genomic observations relate to a tumour phenotype or biological state?
Genomic changes matter because of what they do to the cell. PhenoMap is intended to connect individual molecular observations to the tumour's likely biological state, for example which signalling pathways appear to be altered, so that a list of variants becomes a pathway-level picture.
Main output
Structured phenotype-related observations or interpretation.
What it does not establish
A computationally inferred phenotype is not a measured biological state of the tumour.
Draws on
- Molecular & genomic
- Tumour biology
Output type
- Computational inference
- For treating-oncologist review
Report section
02 Molecular findings
Full layer detailCapability
Understand
Question it addresses
What immune-related signals or biomarkers are relevant to the tumour context?
Some tumour features are relevant to how the immune system interacts with the cancer and to how immunotherapy is evaluated. ImmuneLens is intended to bring the immune-related biomarkers that are available into one context, with their assay and evidence attached.
Main output
Immune-related findings and source and evidence context.
What it does not establish
A biomarker result does not predict that an individual will respond to immunotherapy.
Draws on
- Molecular & genomic
- Immune landscape
- Clinical context
Output type
- Assay-derived finding
- Evidence-linked interpretation
- For treating-oncologist review
Report section
02 Molecular findings · 03 Treatment options
Full layer detailCapability
Connect
Question it addresses
Which alterations may have treatment relevance or warrant further review?
AccessMatch is intended to identify which alterations have evidence connecting them to therapies, whether approved, guideline-referenced or investigational, and to prioritise them for the treating team's review.
Main output
Potentially relevant alterations, evidence and review priorities.
What it does not establish
It does not establish that a therapy will work, that it is approved for the patient's indication, or that it is appropriate for the patient.
Draws on
- Molecular & genomic
- Clinical context
- Evidence ecosystem
Output type
- Evidence-linked interpretation
- For treating-oncologist review
Report section
03 Treatment options
Full layer detailCapability
Explore
Question it addresses
What does the available information suggest about tumour growth dynamics or evolution?
Kinetix is intended to describe what the available information suggests about how the tumour is changing over time. It becomes meaningful mainly when there are repeated measurements, such as serial imaging or repeat molecular testing.
Main output
A structured description of supported dynamics-related evidence.
What it does not establish
It does not forecast how quickly the tumour will grow or when it will progress.
Draws on
- Molecular & genomic
- Imaging integration
- Treatment exposure
Output type
- Computational inference
- For treating-oncologist review
Report section
05 Longitudinal view
Full layer detailCapability
Follow
Question it addresses
What resistance-related mechanisms or alternate pathways may warrant consideration?
Tumours can become less sensitive to a treatment through changes in the drug's target or by switching to alternative signalling routes (bypass pathways). ResistMap is intended to surface findings the evidence associates with resistance, so they can be considered in context.
Main output
Resistance-related signals or hypotheses, and supporting context for clinician review.
What it does not establish
It does not establish that the tumour is resistant to a given therapy.
Draws on
- Molecular & genomic
- Treatment exposure
- Evidence ecosystem
Output type
- Evidence-linked interpretation
- Computational inference
- For treating-oncologist review
Report section
02 Molecular findings · 03 Treatment options
Full layer detailCapability
Explore
Question it addresses
What tumour microenvironment features are represented in the available data?
Tumours grow among blood vessels, connective tissue (stroma) and immune cells, and that surrounding context affects how they behave. This layer is intended to describe what the available data can say about the microenvironment, and how confidently.
Main output
Microenvironment-related observations or interpreted context.
What it does not establish
It does not provide a measured map of the tumour microenvironment unless such measurements were supplied.
Draws on
- Immune landscape
- Tumour biology
- Clinical context
Output type
- Computational inference
- Clinical-context integration
- For treating-oncologist review
Report section
02 Molecular findings
Full layer detailCapability
Connect
Question it addresses
What RNA expression information or signatures are available and relevant?
DNA shows which genes are altered; RNA shows which genes are being actively used. RNA data can support the detection of gene fusions and expression-based signatures. This layer is intended to summarise whatever RNA information the selected pathway provides.
Main output
Expression and signature summaries and their evidence context.
What it does not establish
Without RNA input, this layer produces no expression findings.
Draws on
- Molecular & genomic
Output type
- Assay-derived finding
- Computational inference
- For treating-oncologist review
Report section
02 Molecular findings
Full layer detailCapability
Connect
Question it addresses
How do molecular findings relate to diagnosis, stage, treatment history, comorbidities and intent?
The same molecular finding can mean different things in different cancers, stages and treatment histories. Clinical Correlation is intended to place the findings in the clinical context supplied by the treating team, so that interpretation is case-specific rather than generic.
Main output
A structured interpretation informed by the case context.
What it does not establish
It does not verify clinical information that was not provided, or replace the treating team's judgement.
Draws on
- Clinical context
- Treatment exposure
- Patient factors
Output type
- Clinical-context integration
- For treating-oncologist review
Report section
01 Executive molecular summary · 02 Molecular findings
Full layer detailCapability
Understand
Question it addresses
How can relevant imaging findings or changes contribute to the evolving picture?
Imaging is central to how cancer is staged and followed. Imaging Integration is intended to bring relevant radiology findings, and how they change, into the same view as the molecular results.
Main output
Imaging-related findings and longitudinal comparisons where supported.
What it does not establish
It does not provide an independent radiological read of images.
Draws on
- Imaging integration
- Clinical context
Output type
- Clinical-context integration
- For treating-oncologist review
Report section
01 Executive molecular summary · 05 Longitudinal view
Full layer detailCapability
Connect
Question it addresses
What staging or prognosis-related information is available and applicable to the case?
Staging and risk information shapes almost every treatment conversation. This layer is intended to summarise the staging and prognosis-related information applicable to the case, with the method and its source stated.
Main output
Staging and risk or prognosis context with method and source attribution.
What it does not establish
It does not predict an individual patient's outcome or survival.
Draws on
- Clinical context
- Molecular & genomic
- Imaging integration
Output type
- Evidence-linked interpretation
- Clinical-context integration
- For treating-oncologist review
Report section
01 Executive molecular summary
Full layer detailCapability
Understand
Question it addresses
What response-related evidence or estimates can be reviewed for a possible therapy?
When a therapy is being considered, it helps to see what is known about response in tumours with similar features. This layer is intended to organise response-related evidence and any supported estimates, and to make their limits explicit.
Main output
A response-related assessment with a clear explanation of evidence and uncertainty.
What it does not establish
It does not establish whether this patient will respond to a particular treatment.
Draws on
- Molecular & genomic
- Immune landscape
- Treatment exposure
- Evidence ecosystem
Output type
- Evidence-linked interpretation
- Computational inference
- For treating-oncologist review
Report section
03 Treatment options
Full layer detailCapability
Explore
Question it addresses
What is the starting information set and how is the initial Twin established?
Every Twin starts somewhere. TwinBirth is intended to establish the baseline Twin from the data available at the start of the journey — Day 0 in the six-month journey — and to record what information is missing.
Main output
A baseline structured view, with the inputs used and missing information identified where supported.
What it does not establish
It does not create a complete digital copy of the patient or the cancer.
Draws on
- Molecular & genomic
- Clinical context
- Imaging integration
- Treatment exposure
- Patient factors
Output type
- Twin workflow function
- For treating-oncologist review
Report section
01 Executive molecular summary · 05 Longitudinal view
Full layer detailCapability
Connect
Question it addresses
Which clinical trials may be relevant for further review based on the case and available trial criteria?
Clinical trials can offer access to treatments under investigation, but finding relevant ones is difficult. TrialGraph is intended to identify trials that may be relevant to the case for the treating team to review, with the reasons for each match.
Main output
Potential trial matches and a rationale for review, if the product supports that output.
What it does not establish
It does not confirm that a patient is eligible for, or will be accepted into, any trial.
Draws on
- Molecular & genomic
- Clinical context
- Treatment exposure
- Evidence ecosystem
Output type
- Evidence-linked interpretation
- For treating-oncologist review
Report section
04 Clinical-trial matches
Full layer detailCapability
Explore
Question it addresses
What toxicity or tolerability considerations may be relevant to the patient or therapy?
The best option on paper is not always the right one for a particular person. ToxCast is intended to surface toxicity and tolerability considerations relevant to the patient and the therapies under discussion, including pharmacogenomic (PGx) information where the testing pathway includes it.
Main output
Structured toxicity and tolerability considerations for clinical review.
What it does not establish
It does not determine dosing or predict whether a specific patient will experience a specific side effect.
Draws on
- Patient factors
- Treatment exposure
- Molecular & genomic
Output type
- Evidence-linked interpretation
- Clinical-context integration
- For treating-oncologist review
Report section
03 Treatment options
Full layer detailCapability
Explore
Question it addresses
What treatment-related cost or pathway considerations can be surfaced?
Cost and access shape real treatment decisions. CostPath is intended to make treatment-pathway and cost considerations visible for discussion, with the geography, date and assumptions behind them stated.
Main output
Cost and pathway context with geography, date and assumptions stated.
What it does not establish
It does not provide a quotation, confirm insurance coverage or guarantee availability.
Draws on
- Treatment exposure
- Evidence ecosystem
- Patient factors
Output type
- Evidence-linked interpretation
- For treating-oncologist review
Report section
03 Treatment options
Full layer detailCapability
Explore
Question it addresses
What relevant evidence has changed since the previous review?
Oncology evidence moves quickly: new approvals, updated guidelines, new trials and reclassified variants. EvidencePulse is intended to identify relevant changes in evidence since the previous review, so the treating team can decide whether reassessment is warranted.
Main output
Evidence updates or notifications consistent with the live workflow.
What it does not establish
It does not guarantee that every relevant new publication or approval is captured.
Draws on
- Evidence ecosystem
- Molecular & genomic
Output type
- Evidence-linked interpretation
- Twin workflow function
- For treating-oncologist review
Report section
05 Longitudinal view
Full layer detailCapability
Follow
Question it addresses
What was reviewed or decided, why, and what outcome was subsequently recorded?
Over months, it becomes hard to remember why a decision was made. DecisionLedger is intended to keep a structured record of what was reviewed or decided, the rationale, and what happened next — governed by the treating oncologist's approved access workflow.
Main output
A structured record of review, rationale and outcomes, governed by user permissions.
What it does not establish
It records decisions; it does not make or evaluate them.
Draws on
- Treatment exposure
- Clinical context
Output type
- Structured record
- For treating-oncologist review
Report section
05 Longitudinal view
Full layer detailCapability
Follow
Question it addresses
How does the model change when new data or relevant evidence arrives?
The Twin is designed to change as the cancer journey changes. Dynamic Twin Updates is intended to describe how the Twin is updated when new data or relevant evidence arrives during the six-month journey, and how each change is reviewed and versioned.
Main output
An updated view with change context, where implemented.
What it does not establish
It does not provide continuous, real-time monitoring of the patient.
Draws on
- Molecular & genomic
- Clinical context
- Imaging integration
- Treatment exposure
- Evidence ecosystem
Output type
- Twin workflow function
- For treating-oncologist review
Report section
05 Longitudinal view
Full layer detailCapability
Follow
Five groups
How the layers fit together
01 · Layers 1–6
Six shared core layers
What the tumour's molecular profile contains: clonal structure, origin of variants, mutational patterns, phenotype, immune biomarkers and targetability.
02 · Layers 7–10
Tumour behaviour
How the tumour may behave: growth dynamics, resistance routes, the surrounding microenvironment and expressed genes.
03 · Layers 11–14
Clinical integration
Placing molecular findings in the clinical picture: case context, imaging, staging and response-related evidence.
04 · Layers 15–18
Action and continuity
From analysis to options for review: creating the baseline Twin, trial matching, tolerability and cost considerations.
05 · Layers 19–21
Longitudinal intelligence
Keeping the picture current: evidence monitoring, a record of decisions, and periodic updates to the Twin.
Accessible list
All 21 layers at a glance
- 1CloneTraceClonality and tumour evolution
- 2SourceGuardSomatic and germline interpretation
- 3SigPrintMutational signatures and aetiology
- 4PhenoMapMolecular and cellular phenotype
- 5ImmuneLensImmune landscape and biomarkers
- 6AccessMatchActionable alterations and targetability
- 7KinetixGrowth dynamics and evolutionary trajectory
- 8ResistMapResistance mechanisms and bypass routes
- 9Tumour MicroenvironmentStromal, immune and microenvironmental context
- 10RNA Expression & SignaturesExpression patterns and signatures
- 11Clinical CorrelationClinical-context integration
- 12Imaging IntegrationRadiology and imaging biomarkers
- 13Staging & Prognostic ModelsStage, risk models and outcome estimates
- 14Treatment Response PredictionLikelihood of response and resistance
- 15TwinBirthCreation of the patient's digital twin
- 16TrialGraphClinical trial matching and prioritisation
- 17ToxCastTherapy toxicity risk and tolerability
- 18CostPathTreatment pathways and cost considerations
- 19EvidencePulseContinuous evidence monitoring
- 20DecisionLedgerTracking decisions, rationale and outcomes
- 21Dynamic Twin UpdatesPeriodic model updates over time
- 01 Six shared core layers · Layers 1–6
- 02 Tumour behaviour · Layers 7–10
- 03 Clinical integration · Layers 11–14
- 04 Action and continuity · Layers 15–18
- 05 Longitudinal intelligence · Layers 19–21
