Methodology · dataset v18
How we decide
Every verdict in the checker comes from scientific trigger classes described in our own words — never copied, never a safety certificate.
The lipid rule
Malassezia yeasts are lipid-dependent and cannot make their own fatty acids — they must harvest fats from the skin or from cosmetics. Their enzymes split triglycerides and esters into absorbable fatty acids. So we flag classes the yeast is shown to eat (fatty acids ~C11–C24, fatty esters, polysorbates, ferments), each with a strength of evidence. Nitrogen alone is never enough to flag (the urea lesson): the bar is lipid-substrate evidence.
Evidence levels
Trigger · strong evidence — the class is the yeast's textbook food (e.g. olive oil, the lab growth medium).
Trigger · moderate evidence — solid class membership with direct feeding or formulation studies behind it.
Trigger · limited evidence — honest class-level call (e.g. unlisted plant oils, ferments): most members feed the yeast, this one is not individually verified.
Antifungal · known to help — a separate positive axis for ingredients commonly used against Malassezia, never a claim that a product treats anything.
No known trigger — judged, with an affirmative reason (no C11+ chain, rinse-off, mineral, polymer…).
Not recognized — a typo, or a real ingredient we choose not to judge (usually a botanical extract whose makeup varies). Not judged is not the same as safe.
Limited evidence
Some flags rest on inference — consistency with related ingredients or community experience — rather than on direct feeding data: free fatty alcohols (including cetyl, stearyl and cetearyl), lanolin and wax esters, benzoate/salicylate esters judged by inheritance, ferment-class hits, and "possible oil" ambiguities. They display with limited-evidence labels like every limited call.
The product catalog keeps a zero-flag bar: a non-antifungal product with any flag is not listed (antifungal treatments are the only exception, and show every flag honestly).
Botanical extracts
A plant extract names its source and its form — both matter. An extract made from a non-lipid plant part (leaf, flower, root, stem, bark, herb, or aerial parts) — or from a fruit that does not carry an oil (apple, citrus, berries and the like) — carries no fatty chains, so it is judged no-known-trigger. But extracts from seeds, kernels, nuts, and the oil-bearing fruits (olive, avocado, coconut, palm, sea buckthorn, açaí and kin) stay unjudged unless their solvent is confirmed, because those parts can carry real fats. And anything sold as an oil, oil infusion, macerate, butter, or wax keeps its own (usually flagged) verdict: an oil carrier would have to appear on the label as its own ingredient, and that ingredient is already a trigger — so the product fails either way.
Ethers are not esters
The yeast's documented cutters are ester-cutters: they free fatty acids from ester bonds (that is how it eats Tweens and triglycerides). Ethoxylated fatty alcohols (laureth, trideceth, ceteareth and kin) lock the chain behind an ether bond instead — and no ether-cleavage record exists. So the ether-only forms are judged no-known-trigger, while anything carrying a fatty-acid ester part (polysorbates, PEG stearates) stays flagged, and the free fatty alcohols themselves stay flagged too.
Basis labels
Each new entry carries how its call was made: Peer-reviewed (a fetched, title-matched paper — linked under Sources on its page),Mechanism-based (chain length, ester/amide structure, volatility — reasoning from the lipid rule, no paper claimed), orCommunity-reported (sanity-check against community lists only, never copied).
Limitations
An INCI list alone cannot tell concentration, formulation, or your skin chemistry — so we screen ingredients, never diagnose, treat, or certify products. Dataset rot is real (synonyms, reformulations); entries carry a version stamp and unknown ingredients are researched before any product that depends on them enters the catalog.
Key sources
Theelen et al. 2018 — Malassezia ecology, pathophysiology, and treatment (Med Mycol)
Malassezia are lipid-dependent yeasts — the foundation of the lipid-substrate rule.
The yeast cannot make its own fatty acids, so it must harvest them — from sebum or from cosmetics.
Mayser et al. 1995 — ester hydrolysis depends on the alcohol moiety (Acta Derm Venereol)
Yeast enzymes split fatty esters into edible acids; supports the ester class and informs the fatty-alcohol question.
Ghannoum & Rice 1999 — azoles inhibit ergosterol synthesis (Clin Microbiol Rev)
The yeast builds its own sterols — the basis for not flagging sterols like cholesterol.
Satchell et al. 2002 — 5% tea tree oil dandruff RCT (J Am Acad Dermatol)
The single trial behind the tea-tree helpful-1 call, stated with its limits.
Leeming et al. 1986 — azelaic acid vs P. ovale, little response (Br J Dermatol)
Direct negative record: the reason azelaic stays neutral despite helping the condition.
The review behind free-fatty-acid flags: the yeast harvests C11–24 chains directly.
Brunke & Hube 2006 — MfLIP1 extracellular lipase hydrolyses Tweens (Microbiology)
The yeast secretes fat-splitting enzymes — the mechanism behind fatty-ester, polysorbate and soap flags.
Nenoff & Haustein 2002 — phytosphingosine stimulates Malassezia at low doses (Acta Derm Venereol)
Direct paradox record: lab tests stop the yeast only far above cosmetic doses — lower doses feed it instead.
Not medical advice: we screen ingredient lists for potential triggers — we never diagnose, treat, or certify products. When in doubt, ask a dermatologist.