Database-independent
De novo sequencing reads a peptide straight from its tandem mass spectrum. No reference proteome is required.
- Confidence
- 98%
The AI infrastructure for protein sequencing.
Proteins carry out the work of biology, but reading them is still a bottleneck. Conventional mass spectrometry workflows depend on a reference database. Anything absent from it stays invisible: a novel protein, an unexpected modification, an organism nobody has sequenced.
Seqora reads peptides directly from the spectra, with calibrated confidence attached to every call. InstaNovo is the engine at its core, with the rest of the platform built around it over several years of published work.
Calibrated confidence
Every prediction carries a confidence estimate with FDR control and match metrics, so results can be trusted.
Modification-aware
Post-translational modifications are natively supported. No need to specify bespoke modifications and tolerate result tradeoffs.
Built to run at experiment scale.
The pipeline takes a full run of raw spectra and returns identification tables, with the modification coverage and downstream analysis proteomics researchers actually need.
- Fast
- Thousands of spectra per second
- Accurate
- High identification rates and precision
- 100+
- Modifications supported
- End to end
- Raw spectra through to identification tables
- Downstream
- Analysis workflows supported after identification
From raw spectra to evidence you can inspect.
Seqora is workflow software, not only a model. Projects, runs and results sit in one application, on your own hardware or on a cluster.
01
Acquire
Raw mass spectrometry files are ingested and pre-processed into a run, locally or on HPC.
02
Sequence
Our InstaNovo models read each tandem spectrum and propose the peptide that produced it.
03
Score
Predictions are rescored and calibrated, then filtered at a chosen false discovery rate.
04
Interpret
Results roll up to peptides and proteins, with observed and predicted spectra side by side.
- Confidence
- 97%
- Spectral angle
- 0.93
- Explained intensity
- 84%
- Hyperscore
- 48.2
- Site localisation
- 98.6%
- Ions matched
- 16/20
One platform, many applications.
Standard proteomics
Tryptic discovery workflows, cross-species panels, and label-free quantification.
Antibody & protein sequencing
Reading therapeutic and engineered proteins that have no genomic reference.
Immunopeptidomics
Non-tryptic HLA and MHC peptides, where database search is weakest.
PTM-heavy workflows
Phosphoproteomics and broader modification discovery.
Metaproteomics
Mixed and environmental communities with incomplete reference coverage.
Degradomics
Protease activity and cleavage-product analysis.
Unlocking the dark proteome
We turn proteins into data, unlocking new ways to understand, explore, and work with the building blocks of life.