Clinical Pedigree Drawing — NSGC Conventions, Three-Generation Capture & Risk Models

Short version. Clinical pedigree drawing follows NSGC 2022 conventions on a three-generation-from-the-proband structure, with explicit notation for consanguinity, pregnancy outcomes, twins, adoption, and the proband. The skill is half symbol-set and half drawing fluency; the tool's job is to make the second half cheap so that the first half stays correct.

What "clinical" adds to pedigree drawing

A clinical pedigree records the same family structure as a genealogy diagram but encodes additional structured information: sex assigned at birth (square / circle / diamond), affected status for the catalogued condition (filled or quadrant-filled symbols), age at diagnosis, deceased status with cause and age, partner relationships (horizontal line) versus consanguineous unions (double horizontal line), descent (vertical line from partner line to sibship), sibship birth order, twins (with monozygotic / dizygotic distinction), pregnancy outcomes (live birth, miscarriage, termination, stillbirth) with gestational age, adoption (brackets), carrier status (central dot or half-fill), and proband marker (arrow). The information density is what makes the pedigree useful for inheritance-pattern analysis and for input to family-history risk-model algorithms.

The detail of the symbol set is in the NSGC pedigree notation guide and the pedigree symbols reference. The 2022 update to the conventions, building on Bennett et al. 1995 and the 2008 update, is the current reference.

Three-generation capture in practice

The minimum useful clinical pedigree captures three generations from the proband. The convention is to draw four generations where information allows: I (great-grandparents), II (grandparents and their siblings), III (parents and their siblings), IV (proband and siblings, with cousins represented as offspring of the parents' siblings). The proband's own offspring fill out a fifth generation if relevant. Affected status, age at diagnosis, age at death, and cause of death are recorded for every individual where the information is known; missing information is recorded explicitly (the convention is "?" for unknown affected status and "d." or a vertical bar across the symbol for deceased).

Three-generation capture matters because pattern recognition across two generations is too thin for most analyses. Autosomal dominant transmission is suggested by affected individuals in every generation; autosomal recessive by an affected sibship with unaffected parents (with consanguinity raising the prior probability); X-linked recessive by predominantly affected males inheriting from carrier mothers with no male-to-male transmission. None of these patterns is reliably visible in two generations.

Consanguinity loops

Consanguineous unions appear in pedigrees from communities where cousin marriage is common, in research datasets that include consanguineous families, and in family histories where a remote consanguineous relationship is discovered during ascertainment. The convention is a double horizontal line between the two partners, with the descent line dropping from the lower of the two horizontal lines. The closed loop that consanguinity creates can be drawn explicitly or implied by the partner-line annotation; explicit drawing is preferred when the relationship will be read by anyone other than the original drawer.

Wright's coefficient of inbreeding (F) is computable directly from the structure of a consanguinity loop, and a tool that supports clinical pedigree drawing should compute and display it. F = 1/16 for first-cousin offspring, F = 1/64 for second-cousin offspring, with intermediate values for half-cousin and other configurations. In autosomal recessive disease, consanguinity raises the prior probability of homozygosity at any locus by F times the population homozygosity rate, which is what makes the calculation useful in family-history risk assessment.

Pregnancy outcomes

Recording pregnancy outcomes is part of clinical pedigree drawing because miscarriages, terminations, and stillbirths carry inheritance-pattern information that is lost if they are recorded only as "no living child". The conventions:

Pregnancy: a small triangle (or "P" annotated symbol) for an ongoing pregnancy; gestational age annotated.
Spontaneous abortion / miscarriage: a small filled or unfilled triangle (depending on whether the loss was affected); gestational age annotated.
Termination of pregnancy: a triangle with diagonal slash; the indication (where known: structural anomaly, chromosomal aneuploidy, single-gene diagnosis) annotated.
Stillbirth: standard symbol with diagonal slash; gestational age annotated; affected status recorded if known from post-mortem.
Termination for foetal anomaly (TOPFA): triangle with slash; the specific anomaly annotated.

For families with recurrent pregnancy loss, parental balanced translocation, single-gene disease with high recurrence (for example, osteogenesis imperfecta type II with germline mosaicism), or other inheritance-pattern-relevant features, accurate recording of pregnancy outcomes is part of the family-history record. Detail on individual conditions is in the disease-specific pedigree pages: DMD, Dravet syndrome, Rett syndrome, tuberous sclerosis, achondroplasia.

In-tool risk-model implementations versus canonical computation

The boundary between an in-tool implementation of a published risk-model algorithm (suitable for research, teaching, and exploration) and the canonical clinical-grade computation matters, and Evagene draws it explicitly.

BayesMendel BRCAPRO, MMRpro, PancPRO are implemented in Evagene running on the pedigree. The original BayesMendel R packages developed at Harvard remain the authoritative source. Evagene's outputs are illustrative; clinicians using BayesMendel computations for clinical decisions should use the original packages directly or a clinically-cleared platform that implements them.

Tyrer-Cuzick in Evagene is an IBIS-style approximation of the Tyrer / Duffy / Cuzick 2004 algorithm. The full coefficients of the official IBIS Breast Cancer Risk Evaluator binary are not public, so any in-tool implementation is an approximation. The official binary is the appropriate tool when an authoritative Tyrer-Cuzick computation is needed.

BOADICEA is licensed by the University of Cambridge and is not bundled in Evagene. Evagene exports a ##CanRisk 2.0 pedigree file that the user uploads at canrisk.org. The off-platform pattern is positive evidence of architectural separation between pedigree drawing / documentation and clinical-grade computation, and it is the canonical workflow when BOADICEA output is needed. See CanRisk vs Evagene for the workflow detail.

Family-history scoring tools (Manchester, Amsterdam II, revised Bethesda, NICE CG164 / NG101 illustrative thresholds, Claus 1994, Couch 1997, Frank 2002, Evans 2004, Vasen 1999, Umar 2004, Gail 1989) are implemented as illustrative descriptions of what the published algorithm produces from the family-history pattern. They do not constitute platform recommendations.

Drawing fluency: gesture versus dropdown

A clinical pedigree is often drawn live during a consultation, a teaching session, or a research-data review, where the drawer is also taking the family history at the same time. Fluency matters: a tool that interrupts the conversation to ask which symbol to insert is friction. Evagene's gesture-based drawing mode is designed for this case — partners, parents, siblings, children, and pregnancy outcomes are added by small consistent gestures, and the layout adjusts as the family grows. A traditional dropdown / form mode is also available for users who prefer it.

The drawing mode is independent of the export. Whether the pedigree is drawn by gesture, dropdown, OCR import from a hand-drawn pedigree, or GEDCOM import from another tool, the resulting structured pedigree exports identically.

Standards-grade interoperability

A clinical pedigree should never be locked into a single tool. Evagene supports interchange in:

GEDCOM 5.5.1 — the de facto interchange format; see GEDCOM pedigree software.
PED — for linkage and population-genetics workflows.
Phenopackets v2 — GA4GH-standard phenotype-aware exchange; see phenopackets and pedigree data.
HL7 FHIR FamilyMemberHistory — for downstream consumption by systems that choose to receive structured family-history data.
CanRisk 2.0 — the BOADICEA bridge.
PDF, PNG, SVG — printable artefacts.

Clinical pedigree drawing: NSGC conventions, three-generation capture, and risk models