Evagene for clinical geneticists

If you are a clinical geneticist, your clinic is defined by a small number of repeating shapes: the 45-minute new-patient slot, the 30-minute follow-up, the multidisciplinary team meeting, the phone call to the laboratory, the dictated letter. What ties these shapes together is a single artefact — the family pedigree — which is simultaneously the data collection instrument you use to capture the family, the reasoning tool you use to form a differential, the input to every risk model you run, and the document you send to the referring clinician. You need software that treats the pedigree with the same seriousness you do.

Evagene is built around that observation. It is a browser-based pedigree platform for specialists who work at the intersection of family history, Mendelian genetics, and cancer risk. It supports standard NSGC/ISCN notation, integrates BRCAPRO, MMRpro, and PancPRO, and ships with AI interpretation that uses your own LLM keys. It does not try to replace your EHR or your laboratory system; it aims to be the single place your pedigree lives and moves from.

The clinical geneticist workflow

A new referral arrives — a 42-year-old woman sent by her GP, on the basis that her sister, mother, and maternal aunt have all had breast cancer before the age of fifty. Your administrative team triages the letter against your service's referral criteria and books her into a new-patient slot in six weeks. Before the consultation, a nurse or administrator sends a family history questionnaire; the patient returns it with a hand-drawn tree of her maternal side and a note that she cannot reach her paternal cousins. You have a rough picture before she walks in.

On the day of the consultation, you interview the patient in detail. You confirm the cases she listed, refine the ages at diagnosis, add details she had not thought to volunteer (an ovarian cancer in the maternal grandmother, a male relative with prostate cancer), and record a couple of genetic test results already in the family. The pedigree grows on screen as you talk; you annotate affected status, carrier status where known, deaths, and ages. You also capture consanguinity status, ancestry, and the proband designation.

At the end of intake, you run formal risk assessment. BRCAPRO tells you the proband's probability of carrying a pathogenic BRCA1 or BRCA2 variant, which you compare against your service's testing threshold. You consider MMRpro even though the family is primarily breast cancer, because a prostate cancer plus a single colorectal case lifts it marginally into "worth checking." You run a batch screen across your disease catalogue to make sure nothing else is crossing a threshold you had not thought to check.

You draft a clinical letter: a narrative of the pedigree, the risk model results, the recommendation for testing, and the patient-friendly summary to copy to the patient. You consent for testing, complete the request form, and move to your next patient. Six weeks later, a pathogenic BRCA2 variant is reported. The pedigree updates with the genotype, and you begin the cascade — contacting the proband's first-degree relatives, offering predictive testing, annotating each as testing proceeds.

Every step of that day touches the pedigree. If your software forces you to re-draw it, re-enter family history into a separate risk calculator, or manually copy data into a letter template, friction accumulates across every patient, every clinic, every week. Evagene is designed to remove that friction.

What clinical geneticists need from pedigree software

• Standards-compliant notation — NSGC/ISCN symbols enforced automatically, not offered as an option, so that every pedigree you produce is unambiguous to any colleague who reads it.
• Live-consultation drawing — a canvas that keeps pace with a talking patient, without forcing you to break eye contact to hunt through palettes and dialogs.
• Structured disease annotation — ICD-10 and OMIM codes applied against a curated catalogue, because free text cannot drive risk models or interoperate with downstream systems.
• Integrated risk models — BRCAPRO, MMRpro, PancPRO, and Mendelian inheritance analysis running on the pedigree data you have already entered, with no re-keying.
• AI-assisted drafting — narrative interpretation that respects your house style and uses your own LLM provider, not a vendor-mediated model.
• Multi-audience reporting — clinical letters for the referrer, patient-friendly summaries for families, numerical reports for MDT, all from a single pedigree.
• Cascade-testing support — a pedigree that persists over time, annotates genotype alongside phenotype, and tracks predictive testing status across a growing family.
• Interoperability — GEDCOM, JSON, PDF, SVG, PNG, a REST API, and webhooks so data can move between you, the laboratory, the EHR, and any AI workspace you use.

How Evagene supports clinical geneticists

Gesture drawing during live consultation

Evagene's canvas is built for the way you actually talk to patients. Gesture drawing lets you add a mother, a father, a sibling, and their offspring in a few seconds, without leaving the keyboard or switching palettes. Symbols are applied automatically based on recorded sex; affected status shades the symbol in real time as you annotate a condition; deceased status adds the diagonal line when you record a death. Keyboard shortcuts handle the most common actions — add sibling, add partner, add offspring, mark affected, mark deceased, set proband — so your eyes stay on the patient, not on the interface.

For a new-patient clinic, this usually means that by the time the patient has walked through their maternal and paternal families, the pedigree is complete, standards-compliant, and ready for risk assessment. No post-consultation data entry. No "I'll tidy it up later."

BayesMendel risk models, with no data re-entry

The case for integrated risk models has been made elsewhere — our hereditary cancer risk assessment guide covers the clinical logic in detail. The practical point here is that Evagene runs BRCAPRO, MMRpro, and PancPRO directly against the pedigree you have just built. You do not retype ages, you do not recount affected relatives, you do not set up a separate input form. Select the proband, pick the model, read the carrier probability and lifetime risk. If your service uses thresholds — for example, a 10% BRCA carrier probability as a testing trigger — you see immediately whether you are above or below.

Mendelian inheritance models cover the non-cancer monogenic work. Autosomal dominant, autosomal recessive, and X-linked recessive analyses give expected offspring probabilities for reproductive counselling, and batch risk screening lets you run the entire catalogue against a single proband in one click. That last capability is unusual: rather than asking you to suspect a condition and then run the model, it flags conditions the family's pattern crosses thresholds for. You are less likely to miss the Lynch-shaped family mislabelled as a BRCA question, or the Li-Fraumeni-shaped family labelled as breast cancer.

AI interpretation you can trust

Evagene's AI interpretation engine generates a structured clinical interpretation of the pedigree — key findings, family implications, data gaps, screening recommendations — as a drafting aid. Two design decisions make it suitable for clinical genetics specifically.

First, bring-your-own-key (BYOK) LLMs. You supply your own Anthropic Claude or OpenAI GPT API key, encrypted at rest with Fernet. Interpretation runs against the provider you have already risk-assessed, contracted with, and logged under your own information governance. There is no Evagene-hosted model sitting in the middle of your clinical text.

Second, Analysis Templates. You can author a template that describes the structure and tone your service uses when writing interpretations — the headings, the required sections, the variables to inject (pedigree description, proband name, inheritance pattern, risk results). Every interpretation the service generates follows the same shape, whether it is drafted by the consultant, the registrar, or the counsellor. House style becomes reusable infrastructure.

Clinical reporting in four shapes

A consultation produces several documents: a clinical letter for the referrer and the medical record, a patient-friendly summary, a risk assessment report for the MDT or the testing decision, and an AI interpretation draft. Evagene generates all four from a single pedigree. Reports export as PDF for the correspondence system; the pedigree itself exports as PNG, SVG, or PDF for inclusion in any of them. If you need to hand a colleague a pedigree in GEDCOM to pull into a different system, that option is available too.

Cascade testing and the living pedigree

When a pathogenic variant is confirmed, the pedigree becomes a multi-month tracking document. Evagene annotates genotype alongside phenotype, so a relative who has had predictive testing carries both a carrier status flag and the clinical result. Batch risk screening identifies first- and second-degree relatives who should be offered testing; as each is contacted and tested, you update their record and the pedigree reflects the current state of the family. The consanguinity detection (Wright's coefficient) and karyogram viewer handle the less common cases where related unions or chromosomal abnormalities complicate the picture.

MCP for clinicians who already work in an AI workspace

If you already use Claude Desktop or Claude Code to draft letters, review cases, or triage referrals, Evagene's MCP server exposes 11 pedigree tools — read pedigree, list individuals, add relationship, run risk analysis, generate interpretation, and so on — directly inside that workspace. You can ask your AI assistant to "look at this referral, check the pedigree, and tell me whether this patient crosses our BRCA testing threshold" without switching tabs. The same MCP surface is available to any MCP-compatible client.

A typical clinical geneticist session in Evagene

Step 1. Open a new pedigree. From the dashboard, you create a pedigree for the new-patient clinic slot. Evagene pre-fills proband details from any referral data imported via API or typed in.

Step 2. Build the family live. Using gesture drawing and keyboard shortcuts, you add the proband, her parents, siblings, maternal and paternal grandparents, aunts, uncles, cousins. Affected individuals are annotated with ICD-10 / OMIM disease codes from the 200+ condition catalogue; ages at diagnosis and deaths are entered inline.

Step 3. Run batch risk screening. One click runs every applicable model against the proband. BRCAPRO returns a carrier probability; MMRpro screens Lynch; PancPRO screens pancreatic. The batch surface flags any additional condition where the family crosses a threshold.

Step 4. Generate an AI interpretation draft. You select your service's Analysis Template and issue the interpretation. The draft arrives with headings your service already uses — key findings, inheritance pattern, red flags, screening recommendations. You edit, reject, or accept sections.

Step 5. Produce the clinical report. You export the clinical letter as PDF, the patient-friendly summary as a second PDF, and the pedigree as an SVG to include in both. The reports drop into your correspondence system.

Step 6. Return of results and cascade. Six weeks later, the laboratory returns a pathogenic BRCA2 result. You update the proband's carrier status in Evagene. Batch screening now runs against the family tree identifying first-degree relatives for predictive testing. As each relative is tested, you update their record; Evagene emits a webhook so the laboratory order management system records the cascade in parallel.

Algorithm catalogue

Evagene reproduces twenty peer-reviewed pedigree risk-model algorithms plus a CanRisk / BOADICEA export bridge. Each is an algorithm reproduction for teaching, research, and structured documentation — not a clinical output, not a decision tool, not a recommendation about any individual. What any output means for a specific person is a matter for the published literature and qualified professionals, and is not something this software claims to answer.

The catalogue below lists which published algorithms are implemented and where to read about each. Evagene deliberately makes no statement about when to run which algorithm, or what to do with the output.

No clinical recommendations are made by this software. Each algorithm is an implementation of published literature. What any output means, when to consult it, and what action if any to take, are matters for the published primary literature and qualified professionals — not for this software.

Frequently asked questions

Further reading

• Clinical genetics pedigree tools: a field guide
• Hereditary cancer risk assessment
• Mendelian inheritance calculator
• Pedigree drawing software
• Evagene for genetic counsellors
• Evagene for oncologists
• The Evagene platform