Candida Albicans Lives in Most Healthy Guts, and Scientists Are Finally Mapping It

Bacteria dominate gut microbiome research, but they share that territory with fungi, and scientists are only beginning to understand what a healthy fungal community actually looks like. A study analyzing 317 fecal samples from 147 healthy volunteers in the Human Microbiome Project (HMP) offers the most detailed picture yet of the gut mycobiome, the fungal counterpart to the better-known bacterial microbiome. The findings are both reassuring and surprising: fungi are nearly universal in healthy guts, but their communities behave very differently from bacterial ones.

Saccharomyces, Malassezia, and Candida Dominate the Healthy Gut Fungal Community

Across all 317 samples, the three most prevalent fungal species were familiar names in very unfamiliar roles. Saccharomyces cerevisiae, the same yeast used in bread and beer, turned up in 92.2% of volunteers across all visits, with an even higher detection rate of 96.8% at the sample level. Malassezia restricta, better known as a skin fungus, appeared in 78.3% of volunteers. And Candida albicans, often associated with infections, was detected in 63.6% of healthy participants.

That prevalence of C. albicans in people with no symptoms is notable. As a gut commensal, Candida albicans is capable of adapting to the intestinal environment without causing harm under normal circumstances. The study suggests these three species may constitute a core gut mycobiome, a baseline fungal presence shared across a large swath of the healthy population. Whether their consistent detection reflects true stable colonization or simply repeated exposure through food and the environment remains an open question.

Yeast genera more broadly dominated the samples. Eight of the top 15 most abundant genera were yeasts, and the two dominant fungal phyla were Ascomycota and Basidiomycota, consistent with findings from other body sites.

Why Fungal Gut Communities Are Surprisingly Unstable Over Time

One of the study's most striking findings concerns stability, or rather the lack of it. In bacterial communities, an individual's microbiome tends to resemble itself over time more than it resembles anyone else's. That personal signature is considered a defining feature of the gut bacteriome. The fungal gut community does not follow the same rules.

In this cohort, within-volunteer fungal beta diversity (a measure of how different communities are from each other) was not lower than between-volunteer diversity. In plain terms: your mycobiome at one point in time is no more similar to your own mycobiome months later than it is to a stranger's. The fungal community is highly variable, both across people and within the same person over time.

This instability has real implications. It means the concept of a stable, personalized fungal fingerprint does not appear to hold in healthy individuals, at least not with current sequencing methods. Diet, environment, and transient exposure to fungi in food are all candidate drivers, though the study's metadata did not fully capture these variables.

How ITS2 Sequencing Outperformed Other Methods for Fungal Detection

Characterizing the gut mycobiome is technically harder than profiling bacteria. The study used Internal Transcribed Spacer 2 (ITS2) sequencing as its primary method, a gene region specific to fungi, and compared it against two alternatives: 18S rRNA gene sequencing and mining existing HMP whole-genome shotgun (WGS) metagenomic data.

ITS2 came out ahead. Only 0.01% of HMP metagenomic sequences aligned to fungal genomes, illustrating how easily fungi are swamped by bacterial signal in standard metagenomic approaches. The commonly used ITS1 primers, another popular fungal sequencing target, may also contain mismatches that prevent optimal amplification of certain species, including Malassezia, potentially causing those fungi to be undercounted in earlier studies.

The 18S approach added something ITS2 missed: non-fungal eukaryotes. The single-celled parasite Blastocystis appeared in 25% of volunteers and was associated with increased bacterial diversity, though not fungal diversity. Its presence highlights how much of the eukaryotic gut community remains uncharacterized.

The team also designed new 18S rRNA gene primers (1152F and 1428R) aimed at broader eukaryotic capture while discriminating against bacterial 16S signal, a technical step that could improve future studies.

Fungal Databases and Detection Gaps Still Limit the Field

Even with improved sequencing, identifying what fungi are actually present is constrained by incomplete reference databases. In this study, 17% of fungal operational taxonomic units (OTUs) could only be classified as "Fungi sp.", meaning the sequences did not match any named species in existing databases. Bacterial databases are far more complete, giving bacterial microbiome research a significant head start in interpretive power.

Fungal identification also runs into a biological challenge: dimorphism. Many fungi, including C. albicans, can switch between yeast and filamentous (hyphal) forms, and DNA-based methods cannot distinguish between them or between living and dead cells. That ambiguity matters clinically, because diagnosing invasive Candida albicans infection in the gut remains difficult, and the transition from harmless commensal to pathogen is still poorly understood at the community level.

Additionally, the HMP DNA extraction protocol was optimized for bacteria. The study tested whether a harsher mechanical lysis method changed the picture and found no significant difference in fungal diversity or taxonomy, which is reassuring for data quality, but the low sequencing depth of WGS data likely still undercounts fungal abundance overall.

What Comes Next for Gut Mycobiome Research and Clinical Applications

The correlations between fungal and bacterial populations hint at a structured ecological relationship that future studies will need to untangle. The strongest positive fungal-fungal correlation was between Sarocladium and Fusarium. The most notable negative correlation was between Candida and Saccharomyces, suggesting the two common yeasts may compete for niche space. On the cross-kingdom side, Rikenellaceae bacteria correlated positively with the fungus Botrytis, while Penicillium correlated negatively with the bacterium Faecalibacterium prausnitzii, a species widely associated with gut health.

Understanding these relationships could eventually inform clinical strategies. When C. albicans shifts from commensal to pathogen, the consequences can be severe: invasive candidiasis carries a crude mortality rate of approximately 40 to 55% in critical care settings, often driven by factors like broad-spectrum antibiotics and immunosuppression. Emerging approaches, including nanotechnology-based antifungal therapies designed to overcome drug resistance and toxicity, may eventually benefit from a clearer baseline picture of what a healthy mycobiome looks like before disease sets in.

For now, the HMP mycobiome study establishes that Saccharomyces, Malassezia, and Candida are regular residents of the healthy human gut, not interlopers, and that the fungal community operates by different ecological rules than its bacterial neighbor. Building on that baseline is the next challenge for the field.

Evidence rating

EMERGING — 76/100

Caveats:

• This is a cross-sectional/observational study — it describes who has these fungi present, not what health effects they cause.
• Funding was not disclosed by the authors, so independent verification of the funding source is not possible.
• ITS2 sequencing detects DNA from both live and dead cells and cannot confirm active colonization versus transient environmental or dietary exposure.
• Fungal reference databases remain sparse and poorly curated compared to bacterial databases, meaning some fungi may be misidentified or missed entirely.
• No significant associations were found between mycobiome profiles and host characteristics such as age, sex, or BMI in this healthy cohort.

Large cross-sectional observational study analyzing 317 stool samples from 147 healthy adults using ITS2 sequencing. Establishes prevalence and composition of the healthy gut mycobiome but cannot infer causation or health outcomes. Findings are consistent with smaller prior mycobiome studies but have not yet been independently replicated at this scale.