Reproducible RNA-seq differential expression,
from raw counts to defensible figures.
Upload a count matrix and get DESeq2 differential expression, pathway enrichment, ML-ranked gene targets, and journal-grade figures, each computed in an isolated, single-tenant container with a version-locked methods record. The analysis a core facility would run, reproducible and entirely yours.
Onboarding research labs and pharma teams by request.
Real output
The figures, not a promise of them
A DESeq2 differential expression run on a public Th17 vs iTreg dataset, and the readiness score for a planned design. Interactive on the platform, exportable at journal quality.
Beyond the gene list
From a ranked gene to a costed, bench-ready plan
Most tools stop at differential expression. Prairie ranks your targets, then hands you the reagents, the estimated cost, and the protocol to validate them at the bench. A hypothesis-generation platform, not just a plot.
See the full worked exampleWhat it does
A complete differential expression workflow
Each step uses established, peer-reviewed methods, not a proprietary black box. You get the analysis a core facility would run, without the queue.
Differential Expression
DESeq2-based bulk RNA-seq analysis with volcano plots, PCA, and clustered heatmaps. The same peer-reviewed statistics you would cite in a methods section.
Pathway & GO Enrichment
KEGG pathway and Gene Ontology enrichment via clusterProfiler, rendered as publication-ready dot and bar plots so you can see which biological systems move.
ML Gene Prioritization
Every gene scored 0-100 from statistical confidence, effect size, expression, and network context: uncertainty-weighted, with no hardcoded gene lists.
Publication-Ready Figures
ggplot2 figures and reproducible result tables formatted for journals and grant aims. Export the full package, not a screenshot of a dashboard.
How it works
From count matrix to figures, in four steps
Upload counts
Bring a CSV or TSV count matrix. Sample groups and species are detected automatically, no scripting, no config files.
Run in isolated compute
Each analysis runs in an ephemeral, single-tenant container that is created on demand and destroyed on completion. Clean environment every time.
Explore results
Interactive volcano plots, heatmaps, prioritized gene lists, and filterable data tables. Inspect comparisons without re-running anything.
Export the package
Download figures, result tables, and a version-locked methods record: everything a reviewer or co-author needs to reproduce the work.
Built for rigor
Reproducible science, governed data handling
Built for groups that take data governance seriously: single-tenant per-job compute, role-based access control, version-locked methods, and a deployment model that keeps regulated data inside your own perimeter.
Formal certifications (SOC 2, HIPAA) are not yet in place; the architecture is built to support them, and a DPA is available on request.
Read the security overviewEphemeral isolated compute
Analyses run in per-job containers with no shared state and no cross-tenant access. The environment is torn down when the job finishes.
Your data stays yours
You own every input and output. Data is never used to train models, never sold, and deletable on request. Encrypted in transit and at rest.
Reproducible by design
Version-locked pipelines and deterministic methods mean a re-run returns the same result, and a methods record reviewers can verify.
Built for pharma governance
A bring-your-own-cloud deployment path keeps regulated data inside your own infrastructure and security perimeter.
Deploy and reproduce
Runs in your cloud. Reproduces to the hash.
Two things a regulated buyer asks for: keep the data inside our perimeter, and prove a result is reproducible. Both are built in, not bolted on.
Bring your own cloud
Prairie deploys into your own AWS or GCP account, under your keys and your VPC. Analyses run on your infrastructure; inputs and outputs never leave your security perimeter.
Reproducible to the hash
Every run pins its environment and content-addresses its inputs and outputs. Re-run the same inputs, get the same hash. This is the real record from the run above.
Put your RNA-seq to work today
We are onboarding research labs and pharma teams by request. Tell us about your data and we will get you set up with the right deployment.