• contact@insightvariant.com
Insight Variant · Product overview

AI-Powered Evidence Intelligence for Genomic Variant Analysis

Multi-agent AI platform that classifies genomic variants using ACMG/AMP guidelines. Automatically mines PubMed literature, queries ClinVar, gnomAD, and OMIM, and generates clinical-grade evidence reports for every variant.

How it works

From VCF to clinical-grade classification — in minutes

A managed pipeline that takes a VCF, normalises against GRCh37/GRCh38 with HGVS notation, and returns ACMG/AMP-graded variant reports.

Step 1

Upload VCF

VCF + HPO phenotype terms. GRCh37 / GRCh38 auto-detected. HGVS normalised.

  • VCF, gVCF, joint-called
  • HPO phenotype anchoring
  • Pre-flight QC checks
Step 2

AI Analysis

Six specialised agents annotate, query databases, and mine literature in parallel.

  • Parallel annotation
  • Literature retrieval
  • Functional predictors
Step 3

Evidence Synthesis

Apply ACMG/AMP criteria systematically with per-rule confidence scoring.

  • PVS1 / PS / PM / PP rules
  • Confidence per criterion
  • 5-tier pathogenicity call
Step 4

Clinical Reports

Cancer risk, management recommendations, pharmacogenomic insights — with full provenance.

  • Cancer risk modelling
  • Pharmacogenomic flags
  • Audit trail per variant
Evidence sources

Trusted genomic databases and live literature

Every classification cites the underlying evidence with source, version, and retrieval time.

Genomic databases

  • ClinVar — clinical interpretations
  • gnomAD — population allele frequencies
  • OMIM — gene-disease catalog
  • Ensembl + VEP — transcript-level effect
  • UniProt — protein annotation

Scientific literature

  • PubMed case studies + GWAS
  • Functional cohort data extraction
  • PMID + DOI citations on every claim
  • Study metadata: cohort, effect size, p-value

Standards & formats

  • ACMG / AMP classification framework
  • HGVS nomenclature
  • GRCh37 + GRCh38 references
  • VCF, gVCF, joint-called VCF inputs
  • SIFT, PolyPhen functional predictors
Architecture

A multi-agent pipeline you can audit step by step

Six specialised agents orchestrated by LangGraph — every step, model call, and token logged.

Agents

6 specialised agents

Each agent owns a discrete job: VCF parsing, database annotation, literature retrieval, evidence ranking, ACMG criteria application, report generation.

  • Parallel execution where possible
  • Bounded per-agent prompts
  • Replayable on the same inputs
Orchestration

LangGraph orchestration

Explicit state graph, not a chain of prompts. Branching, retries, and human approval gates are first-class.

  • Deterministic state transitions
  • Failure recovery per node
  • Curator override at any step
Evidence cards

Evidence Intelligence Cards

Every classification is built from cards: PMID + DOI, ACMG criterion badge (PVS1, PS, PM, PP), study metadata, and a relevance score.

  • Source provenance per card
  • ACMG criterion mapping
  • 5-tier pathogenicity rollup
Capabilities

Everything a curator and a bioinformatician need

Eight first-class capabilities — no scripting required.

VCF upload + HPO

Drag-and-drop VCF, attach HPO phenotype terms, get pre-flight QC results.

ACMG 5-tier classification

Pathogenic, Likely Pathogenic, VUS, Likely Benign, Benign — with criteria trace.

PubMed literature mining

Cite-anchored evidence extraction with PMID, DOI, cohort, effect size, p-value.

Gene-Disease Map

30+ curated gene-disease associations with inheritance patterns (AD, AR, XL).

Evidence Cards

Every claim shown as a card with source, version, ACMG badge, and relevance.

Pipeline Monitor

Real-time view of all 6 agents — status, token spend, retry budget, latency.

Unified search

One query across variants, genes, diseases, publications — with filters and saved views.

RESTful APIs

Uploads, variants, evidence, agents, search — automate the workflow end-to-end.

Built for

Four teams. One platform. Different jobs.

Built around how genomic teams actually work — clinicians, researchers, counselors, and bioinformaticians each get their own view.

Clinical Geneticists

Diagnose faster, sign with confidence

  • ACMG classification with criterion-level trace
  • Cancer risk estimates and management recommendations
  • Pharmacogenomic insights tied to the same variant set
  • Cascade-testing guidance for at-risk relatives
Research Scientists

Cohort-scale insight from messy data

  • Literature mining and evidence aggregation per gene
  • Gene-disease explorer with curated associations
  • gnomAD population analysis and allele frequency views
  • SIFT / PolyPhen functional predictions surfaced inline
Genetic Counselors

Plain-language summaries to share

  • Patient-friendly variant summaries auto-generated
  • Inheritance patterns (AD, AR, XL) clearly labeled
  • Evidence-based risk communication scripts
  • External resource links for further reading
Bioinformatics Teams

Automate, monitor, and audit at scale

  • RESTful APIs for batch uploads and headless runs
  • Batch variant analysis across thousands of samples
  • Per-step agent execution tracking
  • Token accounting and model-version audit per call
Security & compliance

Built for clinical environments — not a research notebook

PHI never leaves your perimeter. Every model call is logged, replayable, and auditable.

HIPAA, GDPR, SOC 2

Aligned to clinical-environment requirements out of the box. BAA available. Complete audit trail from upload to signed report.

Genomic data protection

On-premise or private-cloud deployment in your VPC / VNet / project. End-to-end encryption at rest and in transit. PHI never crosses the boundary.

Full audit trail

Every agent step, LLM call, and classification decision is logged with timestamps, token counts, and model versions — replayable on demand.

Ready to transform your variant analysis workflow?

Join leading clinical labs and research institutions. Tell us what you’re trying to do and we’ll set up a hands-on walkthrough within 2 weeks.

Request a demo Contact sales
HIPAA compliant On-premise deployment ACMG/AMP framework Enterprise support