Per-analyte calibration standards have long made quantitative metabolomics testing prohibitively expensive for all but the highest-volume biopharma labs. PyxisLabs, Matterworks, Inc.’s AI-native laboratory, now offers two quantitative metabolomics testing services: Lab analysis at $200 per sample and an enterprise subscription to run the Pyxis co-scientist on a lab’s own liquid chromatography-mass spectrometry (LC-MS) instruments from $25 per sample. Both routes deliver untargeted small-molecule and lipid identification with computational concentration determination, without a physical calibration standard for every analyte.
Key Insight: Per-analyte calibration and manual peak integration have long blocked high-throughput quantitative metabolomics at the contract lab scale. If your lab is repricing or building out metabolomics, lipidomics, or exposomics capacity, the pricing and workflow model announced here is worth a direct comparison against your current cost-per-sample.
Why Does Quantitative Metabolomics Testing Stall at Identification?
Most untargeted metabolomics workflows stop at identification because determining absolute concentrations requires a matched calibration standard, a validated integration method, and significant analyst time for every detected analyte. This requirement makes quantitative coverage impractical at scale, creating a structural split: labs with bounded analyte lists deliver absolute data, while exploratory projects accept relative abundance as the ceiling. On any multi-client project with an open-ended analyte list, the calibration overhead collapses margins before the science even begins.
HRAM LC-MS instrumentation is not the bottleneck. Biology and life sciences testing for biomarker discovery, mechanism of action (MoA) characterization, or exposomics profiling demands broad, untargeted coverage that calibration-standard-dependent workflows cannot deliver at acceptable cost.
How Does PyxisLabs Remove the Per-Analyte Calibration Requirement?
PyxisLabs decouples absolute concentration determination from per-analyte calibration entirely. The Pyxis co-scientist runs on Matterworks’ Large Spectral Model (LSM), a foundation model that interprets raw mass spectrometry data directly and predicts molecular identity and concentration computationally, delivering untargeted coverage and absolute quantitation in a single pass.
Access runs along two paths:
- Lab Services: Ship samples to PyxisLabs. The lab acquires HRAM LC-MS1/MS2 data across hydrophilic interaction liquid chromatography (HILIC), RP, and lipid methods in positive and negative ionization modes, then returns small-molecule and lipid identification, untargeted concentration, and biological interpretation. Priced at $200 per sample, all-in.
- Enterprise/Subscription License: Subscribe to Pyxis as a modular, tiered Model-as-a-Service (MaaS) platform. Pyxis reads your lab’s raw LC or GC-MS data and provides de novo structure identification and quantitation at scale, with SOPs, columns, and universal calibrants included. From $25 per sample, not including instrument operation cost. Small-molecule and lipid modules are available now; the peptide module ships in Q4 2026.
Both models include full data and IP isolation. Pyxis does not train its foundation model on customer data.
| Parameter | Lab Services | Enterprise License |
| Pricing | $200/sample all-in | From $25/sample (instrument costs separate) |
| Instrument Requirement | None (PyxisLabs provides) | Customer’s own HRAM LC-MS/GC-MS |
| Acquisition Methods | HILIC, RP, lipid; positive and negative ionization | Customer-acquired; Pyxis reads raw files |
| Analyte Coverage | Small molecules and lipids (now); peptides Q4 2026 | Small molecules and lipids (now); peptides Q4 2026 |
| Deliverables | Structure ID, untargeted concentration, biological interpretation | Structure ID, untargeted concentration; includes SOPs and calibrants |
| Data Ownership | Full customer retention; no model training on customer data | Full customer retention; no model training on customer data |
| Best Fit | Labs without in-house HRAM LC-MS; episodic high-complexity studies | Labs with existing LC-MS capacity seeking to expand metabolomics throughput |
Labs running pharmacology and drug development work with existing HRAM LC-MS platforms can extend the Pyxis subscription to metabolomics, lipidomics, and exposomics on the same instruments without capital expenditure.
The National Institute for Innovation in Manufacturing Biopharmaceuticals (NIIMBL)-supported validation of Pyxis for bioprocessing metabolite quantitation confirmed that AI-driven absolute quantitation matches the accuracy of traditional LC-MS methods, a reference point any lab director should review before committing to the subscription.
Amy Caudy, PhD, vice president, Biochemical Omics at Matterworks, notes: “A lab can license the modules it needs and immediately apply Pyxis to applications such as biomarker discovery, quantitative MoA, target discovery, lead identification and optimization, or large population studies. When labs need additional capacity, or if they do not have in-house mass spectrometry, they can engage with Lab Services and benefit from the capabilities of Pyxis.”
What Should a Lab Verify before Adopting AI-Driven Metabolomics Interpretation?
Any lab evaluating Pyxis should first confirm the interpretation layer holds up against its existing validated methods for the target application. GLP-compliant preclinical metabolomics, cGMP bioprocess monitoring, and CLIA-compliant clinical metabolomics each require method verification against the lab’s own acceptance criteria before model-driven interpretation replaces a calibration-standard-dependent assay. The SOPs and standardized methods included in the enterprise license provide a starting scaffold for that work.
For non-regulated or early-discovery metabolomics testing, the case is more direct. Labs already running isotope tracing, lipidome profiling, or cellular metabolism studies on LC-MS can apply Pyxis modules to the same raw data and add quantitative output and biological interpretation without new instrumentation.
Which Delivery Model Fits Your Lab’s Metabolomics Workflow?
The $25/sample enterprise pricing and the method verification requirements under GLP, cGMP, and CLIA together define the two parameters any lab must weigh before selecting a delivery model. The lab’s existing regulatory footprint and instrument infrastructure, not the metabolomics application itself, determine both.
This article has been sourced from a press release here and may include content created or refined.
This article was created with the assistance of Generative AI and has undergone editorial review before publishing.