AIVA (AI-assisted Variant Analysis) is an AI-powered genomic diagnostics platform designed for diagnostic labs. It acts as your AI Clinical Analyst, powered by a 4M+ connection knowledge graph (AIVA-KG) to automate variant analysis and deliver genetic diagnoses in 5 minutes instead of weeks. AIVA combines natural language interaction with ACMG-compliant classification to eliminate expert bottlenecks, with zero data retention for complete privacy.

AIVA integrates into your existing genomic workflow after sequencing and secondary analysis. You upload your VCF files from any pipeline, and AIVA performs tertiary analysis. Through natural conversation, you can ask questions, filter variants, and get instant ACMG-compliant classifications with detailed evidence summaries. Clinical reports are generated in your lab's exact format with one click.

Who is AIVA designed for?

AIVA is built for diagnostic laboratories performing genomic testing, including clinical labs, genetic testing providers, and research institutions. It serves lab directors, genetic counselors, variant analysts, clinical specialists, and bioinformaticians who need to analyze genetic variants efficiently while maintaining clinical accuracy and regulatory compliance.

How fast is AIVA compared to traditional genetic analysis?

AIVA delivers sample-to-diagnosis in just 5 minutes. From VCF upload to clinical-grade report, what traditionally takes weeks of expert review is completed in minutes. AIVA can query 5 million variants in under 1 second, and each AI query takes less than 1 minute to search literature, query databases, and filter variants.

How long does it take to generate a clinical report with AIVA?

Clinical reports are generated instantly with AIVA. Traditional report generation takes days because analysts must manually compile findings, gather evidence, and format documentation. AIVA automates this entire process, producing custom clinical reports in your lab's exact format with one click.

How many cases can a specialist handle with AIVA?

Each specialist can handle 5-10x more cases when using AIVA. By reducing per-case analysis time from hours to minutes, AIVA allows your existing team to dramatically increase throughput without sacrificing accuracy or thoroughness. This means labs can scale their testing capacity without proportionally increasing staff.

What is ACMG/AMP variant classification?

ACMG/AMP variant classification is the gold standard framework established by the American College of Medical Genetics and Genomics (ACMG) and the Association for Molecular Pathology (AMP) for interpreting genetic variants. It categorizes variants into five classes: Pathogenic, Likely Pathogenic, Uncertain Significance (VUS), Likely Benign, and Benign. AIVA applies these criteria systematically and automatically, ensuring every classification meets clinical standards.

How accurate is AIVA's variant analysis?

AIVA ensures high accuracy through systematic application of ACMG/AMP criteria, real-time variant-disease association matching, and comprehensive evidence gathering from genomic databases and literature. Every classification includes detailed evidence summaries so analysts can verify the reasoning.

How does phenotype-aware analysis improve results?

Phenotype-aware analysis means AIVA considers the patient's clinical symptoms when classifying variants. Instead of analyzing variants in isolation, AIVA matches variant-disease associations to the patient's specific presentation. AIVA applies specific ACMG codes based on phenotype matching: PP4: Patient's phenotype highly specific for disease. PS4: Prevalence in affected vs controls. PP1/BS4: Cosegregation with disease

How much can labs save using AIVA?

Labs achieve significant cost savings with AIVA by reducing analysis time from weeks to just 5 minutes per case. These savings come from eliminating expert bottlenecks, dramatically increasing throughput without adding staff, and reducing time-to-diagnosis. By automating the most time-intensive parts of variant analysis, AIVA allows labs to process more samples with existing resources.

Why is AIVA more cost-effective than traditional methods?

Each AI query takes less than 1 minute to search literature, query databases, and filter variants. Tasks that previously took an analyst 30-60 minutes of manual searching now happen in seconds. When you multiply this across hundreds of variants per case and dozens of cases per week, the time savings translate directly into cost savings. Your team processes more cases in less time without adding headcount.

What file formats does AIVA accept?

AIVA accepts VCF (Variant Call Format) files, as well as CSV and TSV files if you've already annotated your variants with another tool. Both compressed (.gz) and uncompressed files are supported. With our optional Parabricks pipeline, you can also upload FASTQ files directly from any sequencing platform. You can upload files through the web interface or programmatically via API for automated workflows. AIVA preserves all your custom annotations throughout analysis, whether from VEP, ANNOVAR, SnpEff, or your own tools.

Can AIVA integrate with our existing sequencing pipeline?

Yes, AIVA is pipeline-agnostic. It accepts VCF files from any bioinformatics pipeline including Sarek, GATK, DeepVariant, and Dragen. There's no vendor lock-in and no need to change your existing workflows. AIVA also integrates with Seqera Studio and offers full API access for custom integrations with your laboratory information systems. Alternatively, you can use our built-in Parabricks-powered secondary analysis pipeline to go directly from FASTQ to variants.

Can AIVA process FASTQ files directly?

Yes, with our optional Parabricks-powered secondary analysis pipeline. You can upload FASTQ files from any sequencing technology—both short-read (Illumina) and long-read (PacBio, Oxford Nanopore)—and AIVA will handle the complete workflow from raw reads to clinical report. This eliminates the need for separate bioinformatics infrastructure while still allowing labs with existing pipelines to use their own VCF outputs.

Does AIVA work with WES and WGS data?

Yes, AIVA supports both Whole Exome Sequencing (WES) and Whole Genome Sequencing (WGS) data. The platform is designed to handle the variant volumes from either sequencing approach, providing the same fast, accurate analysis regardless of your sequencing strategy.

How does AIVA handle multi-sample VCF files?

AIVA natively supports multi-sample VCF files. You can upload family trios, case-control studies, or any multi-sample dataset without preprocessing. AIVA maintains sample relationships and allows you to analyze variants across samples while preserving individual sample annotations and quality metrics.

Can AIVA integrate with other AI tools using MCP?

Yes, AIVA supports Model Context Protocol (MCP) integrations in both directions. You can add external MCPs to AIVA to extend its analysis capabilities with your own tools or third-party services. You can also use AIVA as an MCP server in other AI agents like Claude Desktop, allowing you to query variants, search your genomic data, and run analyses directly from your preferred AI assistant.

Is AIVA HIPAA compliant?

Yes, AIVA is built on HIPAA-compliant infrastructure. We have BAA (Business Associate Agreement) agreements with all our cloud providers. The platform is designed from the ground up for healthcare compliance with automatic PHI detection and secure data handling.

Is AIVA GDPR compliant?

Yes, AIVA is fully GDPR compliant for users in the European Union. We process data based on consent, contractual necessity, and legitimate interests. Users have full rights to access, correct, delete, and port their data.

How long do you retain my data?

We practice zero-day data retention for genetic analysis data. Your genetic data is processed in real-time and permanently deleted immediately after analysis is complete. We do not store genetic samples or analysis results on our servers beyond the active session. This approach minimizes privacy risks while ensuring you maintain full control over your data.

Where is my data processed?

AIVA operates in India and the United States. Your data may be transferred to and processed in either location depending on your region. We ensure appropriate safeguards are in place for international transfers in compliance with applicable data protection laws, including GDPR Standard Contractual Clauses where required.

How is patient data protected in AIVA?

AIVA protects patient data through automatic PHI (Protected Health Information) detection to prevent accidental exposure and secure data handling practices. All infrastructure is designed to meet healthcare compliance requirements including HIPAA and GDPR, ensuring patient information remains protected at every step.

What training is needed to use AIVA?

Minimal training is required to use AIVA. The platform uses a natural conversation interface, so if you can describe what you're looking for, AIVA understands. There are no complex interfaces to learn or specialized query languages to master. Analysts can ask questions in plain language and get immediate, actionable results. Most users become productive within their first session.

Can we customize AIVA for our lab's workflow?

Yes, AIVA is highly customizable. You can configure the interface to show only relevant annotations, create custom clinical reports in your lab's exact format, and integrate custom tools using the Model Context Protocol (MCP). With 10+ built-in genomic tools and full API access, AIVA adapts to your workflow rather than forcing you to change how you work.

Solution

Variant Classification & Interpretation

Automated ACMG/AMP-compliant variant classification powered by a 4M+ connection knowledge graph. Evidence-backed, phenotype-aware, and verifiable.

Request Demo Try AIVA Live

Live Demo

Automated Tools Miss What Matters Most: the Patient

Existing automated classification tools apply ACMG criteria in a vacuum. They evaluate variants based on population frequency, computational predictions, and database lookups, but they ignore the patient's clinical presentation. Without phenotype context, critical codes like PP4 (phenotype specificity), PS4 (case-control data), and cosegregation evidence are left on the table.

The result is classifications that are technically compliant but clinically incomplete. Analysts must still manually layer in phenotype-specific evidence, cross-reference gene-disease associations, and adjust classifications case by case. The bottleneck hasn't moved, it's just shifted downstream.

How AIVA Automates Variant Classification

Purpose-built capabilities for variant classification & interpretation

4M+ AIVA-KG Connections

AIVA is powered by a 4M+ connection knowledge graph linking genes, diseases, drugs, and pathways. Each classification includes structured evidence tracking so analysts can verify every applied code and its supporting data.

Phenotype-Aware Classification

Unlike other automated tools, AIVA incorporates the patient's clinical presentation directly into classification. It applies phenotype-specific ACMG codes (PP4, PS4, PP1/BS4) by matching variant-disease associations to the patient's symptoms, producing classifications that reflect clinical context, not just database lookups.

Clinically Benchmarked Accuracy

AIVA achieves 90%+ sensitivity for pathogenic variant detection, benchmarked on 8,387 clinically curated variants from the FDA-approved ClinGen eRepo dataset. Classifications are validated against real-world clinical decisions, not just computational predictions.

Reclassification on Demand

Reclassify any variant for a specific phenotype through natural language. Ask AIVA to re-evaluate a variant with different clinical context and get updated classifications with citations in under a minute.

0%+

Pathogenic Sensitivity

Detection rate for pathogenic variants on FDA-approved ClinGen eRepo data

0M+

AIVA-KG Connections

Gene-disease-drug-pathway knowledge graph powering variant classification

ACMG Codes Covered

Full coverage of all ACMG/AMP evidence codes including phenotype-specific criteria

Ready to Get Started?

Automate Your Classifications

See how AIVA transforms variant classification & interpretation for your lab. Request a demo or try AIVA live today.

Try AIVA Now Contact Us