— SIRIUS Solutions for

Food and beverage safety

SIRIUS helps food chemists identify what migrates from packaging, screen foods and drinks for contaminants no target list anticipates, verify botanical origin, and trace what processing and digestion make of these chemicals — all from a single LC-MS/MS workflow, even when no reference spectrum exists.

Why food, beverage & packaging safety teams use SIRIUS

Food safety methods are built around lists: regulated residues, known migrants, catalogued E&L library entries. But most of what ends up in food was never on a list — non-intentionally added substances from packaging, byproducts of processing, degradation products of residues, and the markers that separate authentic from adulterated.

SIRIUS doesn’t need a matching library entry. It predicts molecular formula, structure and compound class straight from the fragmentation spectrum — and generates candidate structures de novo when no database holds the answer.

Screen

Screen foods, beverages and packaging for known and unexpected contaminants, residues and migrants that sit outside any predefined target list.

Trace

Follow residues and additives through processing, storage and digestion — into the transformation products and conjugates that reveal true exposure.

Verify

Authenticate botanical origin and expose adulteration by comparing chemical fingerprints and elucidating the markers that set samples apart.

FOOD CONTACT MATERIALS & CONTAMINANTS

Identify packaging migrants & contaminants

Plastics, paper, coatings, inks and adhesives release more than their listed additives. Non-intentionally added substances (NIAS) — reaction byproducts, oligomers, breakdown products of antioxidants, impurities in raw materials — were never registered. Add processing aids, migrants from production equipment and flavorings that look like contaminants, and residue methods built on a fixed analyte list miss most of what is actually there.

Non-targeted screening captures it all — as thousands of features per sample, many unanswered by E&L libraries or ambiguous when several entries share one formula. SIRIUS takes you from raw mzML to ranked, confidence-scored annotations: it reads the fragmentation pattern to rank formulas and structures, searches your own E&L or additive inventories alongside public databases, and surfaces the hits worth following up first.

Molecular formula annotation from isotope pattern and fragmentation tree, without needing a library entry.

Molecular structure annotation checks candidates against built-in databases and your in-house E&L lists and additive inventories.

Compound class prediction sorts unknown NIAS into chemical classes before any structure is confirmed.

Spectral library matching against public and in-house reference spectra runs alongside the de novo search.

Field example

In 37 liquors packaged in plastic and glass, about 2,800 features were screened. SIRIUS returned correct formulas for key substances such as DEHP, DEHA and BPA, told apart ethylvanillin and ethylparaben despite their shared formula, and surfaced three compounds outside the study’s original scope.

Field example

Across human milk samples from Canada and South Africa, non-targeted analysis with SIRIUS identified parabens, phthalate metabolites, PFAS, UV filters and synthetic antioxidants.

More SIRIUS discoveries

114 chemicals migrating from food packaging identified, including hazardous plasticizers and persistent non-intentionally added substances.

Koronaiou et al. | J Hazard Mater (2026) →

Fenozan identified as a key biomarker for human exposure to synthetic phenolic antioxidants used in plastics and food.

Meyer et al. | Environ Int (2025) →

DIETARY EXPOSURE & BIOMONITORING

Map what the diet leaves behind in the body

Dietary electrophiles, additives and residues show up in urine, blood and milk as adducts, conjugates and metabolites — molecules that rarely exist in any reference collection. Screening rules built on a single diagnostic neutral loss miss labile species, and public databases cover only a fraction of the adductome. SIRIUS searches purpose-built databases of predicted conjugates and biotransformation products, and still annotates the compounds those databases miss.

Field example

Pairing enzymatic deacetylation with SIRIUS annotation, researchers mapped 847 urinary features to mercapturic acid conjugates — 624 unique structures, 581 of them absent from HMDB and PubChem. After a deep-fried food intervention, 59 conjugates rose significantly.

More SIRIUS discoveries​

Eccrine finger-sweat analysis reveals healthy lifestyle shifts — including normalized cyclamate levels — in children with overweight after nutritional intervention.

Wolf et al. | iScience (2025) →

Interplay between dietary patterns, plasma metabolites and colorectal cancer risk in 680 cases and matched controls.

Bodén et al. | Sci Rep (2024) →

BioTransformer 3.0 integration generates expected biotransformation products for imported structure databases.

Structure database search across custom conjugate libraries, HMDB and PubChem in one run.

Substructure annotation links fragment peaks directly to parts of the candidate conjugate.

De novo structure generation proposes scaffolds for exposure markers that exist in no database.

EMERGING CHEMICAL THREATS

Catch the "forever chemicals" in food packaging

Grease-resistant papers, coated cookware and fluoropolymer processing equipment can all carry PFAS into food. Targeted methods cover a few dozen legacy compounds; the class spans thousands of structures, including replacement chemistries and polymerization byproducts that nobody has catalogued. SIRIUS merges a dedicated PFAS detection strategy with general small molecule annotation, so fluorinated unknowns are flagged and elucidated in the same run as everything else.

Specialized PFAS detection strategy merges PFAS detection with general small molecule annotation

Automated PFAS flagging during LC-MS/MS pre-processing detects strong polyfluorination signatures directly from the data.

Kendrick mass defect plots visualize homologous PFAS series at a glance, making related structures easy to spot as a group.

De novo structure elucidation characterizes novel PFAS candidates alongside well-known compounds like PFOA and PFOS.

Field example

Suspect screening with SIRIUS uncovered a homologous series of at least seven hydrogen-substituted perfluoroalkyl carboxylic acids — byproducts of PTFE polymerization, the fluoropolymer behind many non-stick coatings. Three were confirmed with authentic standards.

More SIRIUS discoveries​

Over 4,000 compounds annotated and hidden contaminants — including PFAS, mycotoxins, pesticides and pyrrolizidine alkaloids — semi-quantified in bee pollen supplements.

Heeren et al. | Anal Chem (2026) →

AUTHENTICITY & BOTANICAL VERIFICATION

Verify origin and expose adulteration

Herbal extracts, spices, supplements and beverages are often substituted with closely related species or cheaper sources that share the same major constituents. What discriminates authentic from adulterated usually lies in minor compounds — the ones with no library spectrum.

With SIRIUS you can compare samples on predicted fingerprints and compound classes before most features are named, then elucidate exactly the markers that separate them.

Molecular fingerprint prediction turns every feature into structural characteristics for sample-wide comparison.

Compound class prediction profiles botanical chemistry without a database lookup.

Field example

Four commercial Spirulina supplements were profiled alongside Chlorella and Amphora, with SIRIUS annotating free fatty acids, polar lipids and pigments. Each microalga showed a distinct chemical signature — glycolipids and porphyrins in Spirulina, phospholipids in Chlorella, fatty acids in Amphora — and the Spirulina products themselves differed markedly depending on cultivation conditions.

More SIRIUS discoveries​

Organic farming increases nitrogen-rich compounds in pea seeds, with cultivar-specific differences in antioxidants and off-flavor compounds.

Kakhkhorov et al. | Crop J (2026) →

Domesticated common bean varieties show reduced metabolic diversity and specialization compared with their wild counterparts.

de Souza et al. | Plant J (2023) →

TRANSFORMATION & PROCESSING CONTAMINANTS

Follow residues and additives through processing

Heating, frying, fermentation and storage rarely leave a molecule unchanged. Pesticide residues degrade, antioxidants oxidize, lipids break down into reactive aldehydes. The resulting transformation products are often more relevant than the parent — and almost never in a database.

SIRIUS expands the search space itself: generate plausible transformation products from a parent compound instead of stopping at the known structure.

Transformation product generation turns a parent compound into a searchable family of likely transformation products.

Apply your own reaction rules — e.g. oxidation or hydrolysis — to extend the search space.

Field example

An impurity-profiling protocol built on SIRIUS molecular formula and structure annotation elucidated 18 impurities — precursors, byproducts and degradation products — 14 of them reported for the first time.

Know how far to trust an annotation

A contaminant finding can trigger a recall, a supplier audit or a regulatory report. Before a candidate goes further, you need to know how much weight it can carry — and be able to show why.

Confidence score estimates the odds that a structure hit is exactly right — similar in spirit to a false discovery rate.

Substructure validation ties fragment peaks to parts of the candidate and highlights shared explanations in analog hits.

Spectral library and analog matching cross-checks against curated public spectra from GNPS, MassBank and MsnLib and your in-house libraries.

From a single sample to surveillance-scale campaigns

Food monitoring means many matrices, many batches and many samples. SIRIUS scales from interactive inspection on a laptop to automated pipelines on a compute cluster — on the same project files.

GUI, CLI and API — one project

Script entire screening workflows, compute in high throughput via CLI or API, then inspect selected results in the GUI. All share the same persistence layer.

Zero-parameter LC-MS import
Import runs directly from mzML or mzXML with built-in feature detection and alignment — or bring features from your preferred pre-processing tool.

Integrates with your tools
Integrations with prominent third-party tools automate data transfer and mapping of identification results.

Custom databases & libraries Build custom structure databases from your target and suspect lists, and add in-house reference spectra for library matching.

See what SIRIUS finds in your data

Millions of queries are processed every year in SIRIUS to characterize the compounds that spectral libraries can’t. Bring your own LC-MS/MS data from food, beverages or packaging and see what surfaces beyond your library.

Related solutions

Environmental and Water Analysis

Pollutants and degradation products in water and soil

PFAS

Known and novel PFAS in environmental and consumer samples

Impurities, E&Ls and Contaminant Screening

Extractables, leachables and degradation products

Metabolism and Biotransformations

Biotransformation routes across biological systems

Pharmaceuticals and Drug Discovery

Drug leads, metabolites and degradants

Agricultural Science and Crop Protection

Biopesticides, pesticide metabolism and impurities

Clinical and Biomedical Research

Biomarker discovery in research samples

Materials Science

Side-products and degradation in polymers

For Research Use Only. Not for use in diagnostic procedures.