Veraxa · Context-aware data ingestion & validation
Validate data against your design and the literature
Upload your data and Veraxa checks it across four layers, file format and quality, consistency with the experiment you designed in Praxis, consistency with what comparable published studies produced (from Lumen), and community-standard conformance, then returns a validation manifest with clear, actionable findings.
checks if a file is well-formed.
links data to what you logged.
checks whether your data matches the experiment you designed and looks like the data comparable published studies produced, before any analysis runs.
CSV/TSV · XLSX · BIOM · QIIME · FASTQ · VCF · SAM/BAM/CRAM · h5ad · MTX · 10x · mzML · FCS · TIFF (≤ 200 MB)
A table of per-sample groups/labels (supplies the design check).
Cross-check data against published studies.
Cross-check against the planned design.
Understanding Veraxa
Veraxa is the platform's data gate. It sits between the experiment you designed and the analysis you're about to run, and answers one question other tools don't: does this data match the experiment you planned, and does it look like the data comparable published studies produced?
What Veraxa is
Most data problems are caught far too late, after the analysis is run, the figures are made, sometimes after submission. Veraxa moves that check to the start. You upload your data, and before any analysis touches it, Veraxa tells you whether it's the data your study was meant to produce, and flags anything that doesn't fit, with a clear, actionable explanation.
What makes that possible is context. Because your design lives in Praxis and the relevant literature lives in Lumen, Veraxa already knows what your data should look like, so it can check the data against your actual experiment, not just against a file specification.
Why it's different
Tells you whether a file is well-formed.
Links your data to the metadata you logged.
Tells you whether your data matches the experiment you designed and resembles what comparable published studies produced, before analysis runs.
That second half, checking data against the body of published work, and doing it together with your own design, is the part no shipped tool does. It's only possible inside a closed loop where the design and the literature are already on hand.
What it checks
Four independent checks, all anchored to one thing, your upload at the centre.
- Format & quality
Is the file readable and sound? Veraxa profiles the contents and surfaces the data-quality issues that matter for your file type.
- Consistency with your design
Does the data match the experiment you planned, the right samples, the right groups, the right balance between them?
- Consistency with the literature
Does the data resemble what comparable published studies produced, and where it differs, which studies set the expectation?
- Privacy & standards
Is the data free of personal identifiers, and does it follow the community standards your collaborators and journals expect?
Each finding comes with what was expected, what was actually found, and what to do about it. Where the evidence isn't strong enough to judge, Veraxa says so rather than guessing, the platform suggests, you decide.
Veraxa never blocks, it warns
Veraxa never silently changes your data and never stops you. It surfaces what it found, ranked by how much it matters, and leaves the call to you. The one firm line is personal data: anything that could identify a person must be removed before the data moves downstream. Everything else is information for your judgement.
What you get
- A clear verdict, proceed, or resolve a few things first, with every finding explained.
- A harmonised, analysis-ready copy of your data (your original is never touched).
- A validation certificate you can keep with your records or attach to a submission.
- A manifest that carries forward to the rest of your workflow.
Supported formats
More formats are added over time. Aligned with the community standards your field already uses (MIxS, MIAME/MINSEQE, BIDS, and standard identifier and taxonomy references).
References & standards
- 1HGNC, HUGO Gene Nomenclature Committee. Approved human gene symbols (genenames.org).
- 2Greengenes 2 and SILVA 138, ribosomal RNA taxonomy reference backbones.
- 3Genomic Standards Consortium, MIxS and MIMARKS minimum-information checklists.
- 4FGED Society, MIAME and MINSEQE reporting guidelines for omics experiments.
- 5BIDS, Brain Imaging Data Structure specification.
- 6WHO GLASS, EARS-Net and WHONET, antimicrobial-resistance surveillance data structures.