Skip to content

Mass spectrometers produce the data. We read it, automatically.

Faster results. The same answer every time. No human error.

Who built it

A problem we lived before we solved it.

Quantivum started in the lab. We spent years reviewing mass-spectrometry results by hand and saw the same problem everywhere: skilled people spending their days on peaks.

We are scientists who learned to build software. Our founder, Peter Kubiniok, spent fifteen years on instruments; our machine-learning expertise was built in partnership with Mila, the Montreal AI institute.

Working with

  • McGill University
  • Yale University
  • MRM Proteomics Inc.
  • Academia Sinica, Taiwan
  • Mila — Quebec AI Institute

Completed projects

Built for labs. Published when it mattered.

Product · Targeted LC/GC-MS

MRMpipe

Automated peak detection, deconvolution and QC scoring for MRM and PRM assays.

Product page
Problem
Hours of manual peak review per plate, operator-dependent.
Built
Detection, deconvolution, cross-sample QC and a review score, with a web interface.
Outcome
Three days of review become minutes, same rules for every sample.

Database · Proteomics

SysQuan

Searchable peptide-panel database with QC review and automated transition-list export.

sysquan.com
Problem
Peptide panels scattered across spreadsheets, no QC record.
Built
Searchable database with QC review, roles and transition-list export.
Outcome
One source of truth; lookups are instant.

Machine learning · Immunopeptidomics

MHCvalidator

ML rescoring engine for peptide identification.

Nature Communications, 2024
Problem
Conventional scores could not separate weak identifications from false positives.
Built
A rescoring model on domain features with FDR control built in.
Outcome
More true peptides at the same FDR; in use in immunopeptidomics labs.

Dashboard · Quality control

MhcVizPipe

Interactive QC reporting for immunopeptidomics experiments.

Mol. Cell. Proteomics, 2021
Problem
No quick read on allele coverage, contamination and reproducibility.
Built
Pipeline plus dashboard: a one-page QC report per experiment.
Outcome
A day of scripting becomes a glance, comparable across labs.

R package · CRAN

RHybridFinder

Automated detection of hybrid HLA ligands.

On CRAN
Problem
Hybrid peptides needed manual cross-referencing against the proteome.
Built
The scan as a tested R package with a vignette, on CRAN.
Outcome
Installable in one line; reproducible without bespoke scripts.

Publications

More than 13 peer-reviewed publications.

Selected. Complete list on PubMed.

Try it on your own batch.

Up to 100 samples. Full report. Free.