Fern Gene | Platform Architecture

Platform Architecture

Five layers already working together. Each layer has real product capabilities today — this is not a roadmap.

Layer 1

Insurance

Insurance advisors serve as family account managers. They help route clients to the right next step — testing, import, or family setup — without ever accessing raw genomic data.

•Family account management through advisor channel
•Client routing: has data / needs testing / still comparing
•Insurance product review aligned to life stage (maternity, critical illness, long-term care)
•Compliance boundary enforcement — no genomic data in underwriting
•Long-cycle client relationship tracking

Layer 2

Genomics

1,185 production rules across 17 areas. Local-first analysis with no cloud upload. Supports VCF, gVCF, BAM, CRAM, and common consumer formats.

•1,185 production rules across 17 interpretation areas
•Local-first import and analysis — no raw file upload
•Supported formats: txt, VCF, gVCF, BAM, CRAM, .genomesprofile
•Pharmacogenomics: 150+ rules covering metabolism and dosage
•Hereditary disease: 471 rules covering carrier and family risk
•Evidence Room with sourced, checked references per result

Layer 3

Medical

Integration with annual checkups, lab follow-up, and pharmacogenomics for real treatment decisions. Not a diagnostic — a bridge to more informed clinical conversations.

•Annual checkup integration — lipid, liver, glucose trends over time
•Pharmacogenomics for anesthesia, pain management, and chronic medication
•Lab follow-up with genomic context (44 production rules)
•Specialist referral signals for hereditary cancer, FH, and sudden-death risk
•Clear boundary: personal interpretation, not clinical diagnosis

Layer 4

Elderly Care

Chronic disease baselines, multi-drug pharmacogenomics, and home care entry points for the longevity operations stage.

•Chronic disease baseline tracking for adults 55+
•Multi-drug pharmacogenomics for polypharmacy management
•Home care service entry points and coordination
•Long-term care insurance review aligned to health trajectory
•Family caregiver coordination and status sharing

Layer 5

AI

On-device AI drives rule execution, result prioritization, and research. Genomics, RNA, and protein research pipelines are active. Research outputs are checked before becoming user-facing results.

•On-device rule engine with 1,185 production rules
•Result prioritization across all 17 areas
•Active genomics research pipeline with real sample validation
•RNA and protein research lanes in progress
•Research-to-product gate: checked evidence only, never raw research outputs

Technical architecture

Local-first runtime

Import, analysis, and storage run on-device. The encrypted local profile is the primary data object — not a cloud record.

Compact encrypted profiles

Long-lived storage uses compact encrypted profiles, not raw files. Heavy formats like BAM and CRAM are processed and compacted.

Minimal online surface

Partner review, support diagnostics, and entitlement sync use minimal status metadata. No VCF, gVCF, BAM, or CRAM data leaves the device.

Research metrics

17 interpretation areas already live

1,185 production rules in the engine

1,116 content points shipped

10 checked evidence sources

Real-sample validated research pipeline

See how it works For families