74%
Roughly 74% of wet-lab scientists have no programming experience
MVP · early access
An agentic bioinformatics analyst: it runs the analysis, fixes what breaks, and proves the result.
Contig takes raw sequencing data all the way to a verified, reproducible answer: it selects and runs the right pipeline on your own compute, diagnoses and self-heals failures, verifies the output, and writes a portable record anyone can reproduce.
# Python, any OS with Python 3.12+
pipx install contig
# run → self-heal → verify → reproduce
contig run --run-id smoke
# verdict + provenance + repair chain
contig show smoke
No data of your own needed — the smoke run uses nf-core/rnaseq's bundled test profile. From a source checkout, prefix each command with uv run.
01 Why
Producing an end-to-end analysis from raw sequencing data needs a rare pairing of domain biology and computational engineering.
74%
Roughly 74% of wet-lab scientists have no programming experience
~17%
frontier models still reach only ~17% on real bioinformatics analysis
Turning English into a workflow is the easy, crowded part. Running it, fixing it, and proving it is the unsolved part, and that's Contig.
02 What it does
Describe your goal in plain language; Contig proposes a curated pipeline and params to approve, and flags problems (no replicates, single-end, missing reference) before a multi-hour run.
When a step fails (OOM, a missing index it can build, version mismatch, malformed input), Contig diagnoses it, applies a safe patch and retries, showing you the repair chain. A risky fix pauses the run for your approval instead of guessing.
Every run ends in an honest verdict — PASS, WARN, FAIL or UNVERIFIED — backed by real checks, never a bare “done”. contig show --explain names the exact checks that drove it.
Each run pins inputs (checksums), pipeline revision, params and tool versions into a portable record. Your reads never leave your machine; only hashes are recorded.
03 A closer look
04 The verdict
Every run ends in one of four verdicts, drawn from QC-metric thresholds, structural and integrity checks on the output files, cross-tool concordance, and biological-plausibility checks.
Ran to completion and every QC check passed.
Completed, but a QC check is borderline; look before you trust it.
A task or QC check failed; do not trust the output.
Completed, but nothing checked it, so we won't claim it's correct.
Verification is scoped honestly per assay and never over-claims: a check earns a
FAILband only where a grossly-wrong value is distinguishable from unusual-but-real science — so an empty call setFAILs, while a low-purity tumour's VAF stays aWARN, andUNVERIFIEDnever masquerades as either.
05 Status & limits
The Layer-2 core — run → capture → self-heal → verify → reproduce — works end-to-end on nf-core/rnaseq, as a CLI and a local dashboard. The other assays ship a curated pipeline plus an assay-aware QC pack and are being hardened toward the same bar.
nf-core/rnaseq is exercised end-to-end, including the self-heal loop, in CI. For other assays the checks are scoped to what that assay makes distinguishable — no verdict claims more than its checks support.
One command lands unchanged on Docker locally or AWS Batch in your own account, via Nextflow's native executors. The dashboard is localhost-only today.
Apache License 2.0 — permissive, with an explicit patent and trademark grant. Use, modify and redistribute Contig, including commercially, under its terms. Third-party pipelines keep their own licenses.