From sequencing data to a signed report.
One platform.
Helixen runs WGS, WES and panel analysis on GRCh38, annotates findings against ClinVar, gnomAD, COSMIC and PharmGKB, and hands the ordering physician a structured draft with the evidence attached. We build the software — the laboratory and its physicians own the clinical decision.
- Sequencing types supported
- WGS · WES · Panels
- Reference build
- GRCh38
- GATK · VEP · CNVkit · Manta
- Open toolchain
- Where compute runs
- Your infra or ours
Orders
| Date | Analysis type | Patient | Status | Progress |
|---|---|---|---|---|
| 12/05/2026 | Oncology — WGS (tissue) | A. Müller | In progress | 64% |
| 11/05/2026 | Hereditary — WES (blood) | P. Novák | In review | — |
| 10/05/2026 | General screening — AMP (saliva) | L. Costa | Completed | 100% |
| 09/05/2026 | Oncology — Panel (tumor/normal) | J. Park | Awaiting payment | — |
| 08/05/2026 | Hereditary — WGS (plasma) | M. Rossi | Completed | 100% |
The workflows a genomic lab actually runs
Oncology, hereditary and screening cases go through the same pipelines, the same interpretation workspace and the same audit trail.
Oncology
Somatic profiling for tissue and liquid biopsy: driver mutations, tumor CNV, therapy-associated variants from COSMIC and curated evidence.
Coverage
Somatic WGS/WES and targeted panels, including ultra-deep amplicon runs.
Hereditary
Germline analysis with ACMG-oriented classification criteria, carrier status, CNV and structural variants, repeat expansions and pharmacogenomics.
Coverage
Germline WGS/WES for blood and other germline material.
Screening and research
Broad genomic characterization for preventive programmes and research cohorts, with the same pipelines and the same audit trail.
Coverage
Configurable panels and workflow templates per programme.
One platform, the full genomic picture
Not a single test — the pipelines, the evidence, the review workspace and the reporting, connected end to end.
Variant calling
Germline calling with GATK HaplotypeCaller, somatic with Mutect2, on GRCh38. Every run records its tool versions and parameters.
CNV and structural variants
CNVkit for copy number, Manta for structural rearrangements, ExpansionHunter for repeat expansions — findings single-variant tests miss.
Pharmacogenomics
Drug–gene annotation from PharmGKB and CPIC guidance, reported alongside the germline findings.
Sample identity and kinship
somalier-based relatedness and identity checks catch sample swaps before a report leaves the lab.
Interpretation workspace
The reviewing physician inspects every finding down to the reads: genome browser, coverage, gene and protein views, evidence links.
Orders and audit trail
Order lifecycle, role-based access, per-order audit log and versioned reports — who changed what, and when.
Reference build, tooling and evidence sources
How a case moves through the platform
Four stages, each visible to the lab in real time — from the order to the signature.
The lab creates the order
A laboratory or clinic account registers the case, the material and the workflow. Patients do not order directly.
Sequencing data goes in
FASTQ uploads in resumable chunks, or the platform picks the run up from your storage.
Pipelines run, tracked live
Alignment, calling, CNV/SV, annotation and QC — each stage visible with its logs and metrics.
The physician reviews and signs
A structured draft with traceable evidence goes to the ordering physician, who edits, approves and signs it off.
A report the reviewer can defend
No raw data dumps and no unsourced statements. The platform assembles the findings, links the evidence behind each one, and keeps every revision — so the physician who signs it can show where each line came from.
- Structured findings with the evidence behind each one linked
- Clinically relevant variants prioritized, the rest kept in context
- Reviewed, approved and signed by the ordering physician
- Versioned: every revision of a report is preserved
- Machine-readable exports alongside the document (VCF, CSV)
Where the AI helps — and where it stops
Writing up a genome takes hours of a specialist's time. The model drafts that text from the structured results and shortens the write-up. It does not diagnose, and it does not sign.
A draft, not a decision
The model turns structured results into readable narrative text. It reaches nobody until the reviewing physician edits and approves it.
Traceable to the data
Every statement in the draft points back to the finding and the source it came from, so the reviewer can check it instead of trusting it.
Machine-generated text is a draft for a qualified specialist. It is not a diagnosis, not a treatment recommendation, and must not be used as one. The reviewing physician of the ordering organization approves and signs every report.
Genomic data, handled like genomic data
What we can state precisely, we state precisely — and we say where the boundary of our control is.
Role-based access
Client, physician, laboratory and administrator roles; each order is visible only to the people attached to it.
Encrypted in transit
All traffic runs over TLS, and integration credentials are stored encrypted, not in plain configuration.
Data stays where you want it
Run compute inside your own perimeter as an external worker, or let the platform provision ephemeral machines per run.
Audit trail
Order events, report revisions and sign-offs are recorded — the history is there when someone asks.
For laboratories and clinics
Helixen is deployed for the organization that runs the testing. Your laboratory keeps the accreditation, the samples and the clinical responsibility; the platform carries the pipelines, the evidence and the paper trail.
Partner laboratories and their physicians sign in with accounts issued by their administrator.
- Bring your own compute or use ephemeral cloud machines per run
- Configure workflows and panels per programme
- White-label reporting for your own brand
- One workspace for oncology, hereditary and screening cases
Frequently asked questions
What the platform does, what it does not do, and who is responsible for what.
Who is Helixen for?+
Genomic laboratories, clinics and research groups. Helixen is the software layer between sequencing output and a signed report: pipelines, annotation, an interpretation workspace and reporting. We do not operate a laboratory and do not accept samples from private individuals.
What does the platform actually run?+
WGS, WES and targeted panels aligned to GRCh38. Reads are processed with fastp, BWA and samtools; germline variants with GATK HaplotypeCaller, somatic with Mutect2; CNVs with CNVkit, structural variants with Manta, repeat expansions with ExpansionHunter and sample identity with somalier. Annotation runs through Ensembl VEP against ClinVar, gnomAD, COSMIC, PharmGKB and OMIM.
Who signs the report?+
The qualified physician of the ordering laboratory or clinic. The platform prepares a structured draft with traceable evidence, records who reviewed and approved it, and versions every change. Helixen does not employ physicians and does not issue clinical conclusions on its own behalf.
Where does the AI fit in?+
The language model drafts narrative text from the structured analysis results and is never the final word: nothing reaches a report until the reviewing physician edits and approves it. Machine-generated text is not a diagnosis and is not intended to be used as one.
Where does the analysis run?+
On your infrastructure or ours. Compute nodes attach as external workers over a private link, or the platform provisions ephemeral cloud machines per run and tears them down afterwards. Sequencing data can stay inside your perimeter.
What about validation and accreditation?+
Analytical validation against reference materials (GIAB) is in progress, and we publish results as they are measured rather than claiming performance we have not verified. Helixen is not CE-IVD marked and is not registered as a medical device today; the platform is used under the responsibility of the ordering organization.
Run your genomic analysis on Helixen
See what the platform covers today, or sign in to your laboratory workspace.