Evagene for research geneticists

Research genetics sits at a different point on the workflow spectrum from clinical practice. Your cases arrive as cohorts rather than one-off consultations; your outputs are figures, tables, and interpretive statements that support a hypothesis rather than letters that support a patient's care plan; your pedigrees are often inherited from a collaborator's file, not drawn from scratch while talking to a proband. Your software requirements reflect that: you need programmatic access, reproducibility, data portability, and publication-quality visual output at least as much as you need a fast drawing canvas.

Evagene is built as a platform as well as an application. Underneath the browser UI sit a scoped REST API, HMAC-signed webhooks, an MCP server for AI-agent integration, and a full suite of imports and exports. For a research geneticist, this means you can script against Evagene the same way you script against Biopython or GATK — build a cohort, run analyses across it, persist outputs, and reproduce the pipeline when a reviewer asks. The application remains available when you want a graphical view.

The research geneticist workflow

A typical research genetics project begins with a cohort: a family-based study of a rare disease, a population cohort with consented genealogy linking, a retrospective audit of clinically-captured pedigrees, or an ascertainment of extended kindreds from a biobank. The pedigrees arrive as files — GEDCOM from a genealogical source, a custom JSON export from a clinical system, PED files from a linkage analysis, occasionally images that require OCR. Your first task is to get them into a form you can work with.

Then comes analysis. Depending on the question, this might be inheritance pattern classification across the cohort, carrier probability estimation using BayesMendel models, batch risk screening for specific conditions, identification of consanguinity, or extraction of standardised phenotype annotations for downstream genotype-phenotype association analysis. You want to run these analyses across hundreds or thousands of pedigrees, not one at a time. You want the results in tabular form so you can join them to your genotype data or statistical models.

At the end of the project, you need outputs. Publication-quality pedigree figures for the manuscript, structured data for the supplementary tables, standardised interpretations that can be audited by collaborators and reproduced by reviewers, and a data deposit in a format that will survive beyond the life of the project.

Every step of that pipeline benefits from programmatic access, file format portability, and reproducibility of interpretation.

What research geneticists need from pedigree software

• Bulk import from GEDCOM, JSON, PED (via GEDCOM conversion), 23andMe, and pedigree images.
• Programmatic API — scripted pedigree creation, modification, analysis, and export.
• Batch analysis — BRCAPRO, MMRpro, PancPRO, Mendelian inheritance, and batch risk screening across a cohort.
• Publication-quality export — SVG (vector, editable), PNG, and PDF.
• Reproducible interpretation — parameterised templates that produce consistent outputs across cases.
• Data portability — export in the same standards used for import, so the pipeline is not locked in.
• Agentic AI integration — MCP tools so that AI assistants can be part of the research workflow without custom glue code.
• Standard disease annotation — ICD-10 / OMIM coding that interoperates with downstream analytical tools.

How Evagene supports research geneticists

Bulk import from multiple formats

Evagene imports GEDCOM 5.5.1 (the de facto standard for pedigree exchange), JSON (for cleanly-structured modern data), 23andMe exports (for cohorts with direct-to-consumer genotype and trait data), XEG (a legacy pedigree format still in use in some settings), and pedigree images via OCR (for archival cohorts where only a scanned figure survives). For PED and LINKAGE-style tabular inputs, the recommended route is to convert to GEDCOM using any of the widely-available open-source converters, then import. A detailed discussion of GEDCOM in a genetics context lives in our GEDCOM pedigree software guide.

For a study with 500 pedigrees, you might loop over GEDCOM files in a directory, call the REST API import endpoint for each, and have the cohort loaded into Evagene in minutes. Imports are idempotent at the pedigree level; re-running an import against the same file updates rather than duplicates.

REST API and MCP for scripting

The REST API covers the full lifecycle of pedigree data. Endpoints exist for pedigree CRUD, individual CRUD, relationships, disease annotations, 23andMe genotype import, risk analysis (all BayesMendel models and Mendelian inheritance), batch screening, report generation, and export. API keys are in the format evg_ followed by 43 characters, SHA-256 hashed at rest, with read, write, and analyse scopes. Rate limiting prevents runaway scripts from destabilising your own instance. Webhooks signed with HMAC-SHA256 fire on eight event types — pedigree and individual CRUD, analysis completion, import completion — so a research pipeline can react to events rather than polling.

The MCP server is an alternative route into the same capabilities. If your research workflow already involves an AI assistant in Claude Desktop or Claude Code, the MCP tools (11 of them) let the assistant read pedigrees, add individuals, run analysis, and generate reports directly. For exploratory cohort analysis — "find me the families in this dataset that cross the BRCA testing threshold and summarise their pedigree patterns" — this is substantially faster than writing the equivalent script.

Batch analysis across cohorts

Risk analysis in Evagene runs on the pedigree data without re-entry. For a single pedigree this is routine; for a cohort, it is the difference between an afternoon and a month. Scripting against the REST API, you loop over the cohort, invoke the batch screen endpoint on each proband, and persist the structured JSON results — carrier probabilities, lifetime cancer risks, threshold flags — to a research database alongside the genotype data. The Mendelian inheritance analysis endpoint does the equivalent for monogenic conditions.

Consanguinity detection via Wright's coefficient runs similarly across the cohort: a single endpoint returns the coefficient of inbreeding for any specified individual, allowing you to flag consanguineous unions for subsequent analysis.

Publication-quality figures

Pedigrees export as SVG, PNG, and PDF. The SVG output is vector-based, editable in Inkscape, Illustrator, or any text editor for surgical tweaks before submission. Symbols follow NSGC/ISCN nomenclature — squares, circles, diamonds, shaded affected, diagonal-line deceased, proband arrow — so a reviewer in any subfield of genetics reads the figure correctly. For manuscript figures that need bespoke legends or annotations, the SVG export is the starting point; exporting to PDF produces a cleaner drop-in for LaTeX or Word.

Reproducible interpretation via Analysis Templates

Analysis Templates are parameterised AI prompts with variable injection — placeholders such as {{pedigree_description}}, {{proband_name}}, {{inheritance_pattern}}, {{risk_results}} are replaced with concrete values at runtime. A research group can author a template that codifies its interpretation methodology (how to describe the pedigree, which risk thresholds to emphasise, what caveats to include) and apply it across every pedigree in the cohort. The output is structurally consistent, reviewable, and re-runnable. For a reviewer's reproducibility request, you supply the template, the inputs, and the resulting interpretations; re-running the same template against the same inputs produces the same output (within the determinism of your chosen LLM provider).

Because Evagene uses a bring-your-own-key model for LLMs, the language model used for interpretation is one your lab has already procured and can document. Encrypted at rest with Fernet; requests go directly to the provider.

Data portability and deposit

Everything you put into Evagene can be extracted. Pedigrees export as GEDCOM 5.5.1 (vendor-neutral, archival-quality), JSON (modern structured), PNG, SVG, and PDF. Reports export as PDF. The REST API returns the same data in JSON form for programmatic access. For a journal requiring supplementary data deposit, GEDCOM is typically the right format; for an institutional repository, a bundle of GEDCOM plus JSON covers both human and machine access.

A typical research geneticist session in Evagene

Step 1. Bulk import the cohort. A script loops over GEDCOM files, calls the pedigree import endpoint on each, and catches the returned pedigree IDs.

Step 2. Run batch analyses. For each pedigree, the script invokes risk analysis endpoints — BRCAPRO, MMRpro, PancPRO, batch screen, inheritance analysis — and persists the JSON outputs into a research database joined to the genotype data.

Step 3. Apply an Analysis Template. The lab's interpretation template runs against each pedigree; the structured outputs are stored alongside the numerical risk results for subsequent review.

Step 4. Export publication figures. For the kindreds featured in the manuscript, the script exports SVG pedigrees; a graphics tool polishes them for submission.

Step 5. Deposit. The cohort exports as GEDCOM plus JSON for the journal's supplementary data; the interpretation templates and analysis scripts are versioned in the lab's repository alongside.

Frequently asked questions

Further reading

• GEDCOM pedigree software
• Mendelian inheritance calculator
• The Evagene platform (API, MCP, webhooks)
• Hereditary cancer risk assessment
• Evagene for clinical geneticists
• Evagene for educators and students