
A look back at MMHP's 2019 million-sample plan, its research ambitions and why the launch did not establish personal diet or testing guidance.

In a report dated November 8, 2019, BioSpectrum Asia described the launch of the Million Microbiome of Humans Project (MMHP), an international initiative with a goal of sequencing and analyzing one million microbial samples. The project was announced at the 14th International Conference on Genomics and proposed a database covering several body sites and health conditions.
The scientific starting point in the launch report was the need for a broader research foundation. Its leaders wanted to compare microbial patterns across populations and health conditions, with potential medical applications further ahead. That is different from announcing that a particular microbial pattern defines good health or that a particular diet produces it.
In that sense, MMHP reflected an evidence-building approach rather than a change in nutritional consensus. BioSpectrum Asia quoted Karolinska Institutet's Lars Engstrand describing the proposed map as a data foundation for microbiome research. The emphasis was on creating a resource from which later studies could work.
The project's scope also matters. Its planned samples included the intestine and mouth as well as the skin and reproductive tract. This was a human microbiome initiative with a wider brief than gut health alone. Findings from the proposed resource would therefore need to be interpreted in relation to the body site being studied.
For a reader following broader microbiome science, the useful distinction is between building a map and establishing what a person should do. A map organizes observations. A recommendation requires evidence relevant to a choice and its outcome. The launch belonged to the first part of that process.
According to BioSpectrum Asia, the project's stated goal was to sequence and analyze one million samples over three to five years. The scale was central to the proposal, but so was its breadth across populations and body sites. The proposed change was a larger basis for studying human microbial patterns.
The launch report listed founding partners including Karolinska Institutet and the University of Copenhagen. Technical University of Denmark and Shenzhen BGI Research were also named. Other listed partners included MetaGenoPolis at France's INRA and the Latvian Biomedical Research and Study Centre, alongside Shanghai National Clinical Research Center for Metabolic Diseases.
BioSpectrum Asia also reported that partners had completed an initial 10,000 metagenomic samples. The report said BGI Research would provide research platforms and analytical resources for 200,000 samples. These were different parts of the announcement: an early milestone and a commitment of support toward the larger project.
The intended output was a human microbiome map and a large database spanning different populations and health conditions. Project leaders presented this as research infrastructure. In plain language, that means an organized body of information intended to support future scientific questions rather than a finished answer to each one.
For example, the research ambition included comparing patterns in health and disease. BioSpectrum Asia reported that Shanghai researcher Liu Ruixin also discussed comparisons before and after treatment as a possible route toward future therapies. That description sets out a direction for research. The proposed value was in making those comparisons possible at a broader scale.
A research reference database and a clinical reference range serve different purposes. The database can be material for comparing the groups represented in it. A clinical range would go further by establishing how an individual's result should be interpreted. Keeping those concepts separate helps prevent the word “reference” from sounding like a personal health score.
This also clarifies how to read discussion of the search for a best diet. The scientific question is not simply whether researchers can describe microbial differences. It is whether a specific dietary choice produces a meaningful benefit in a relevant group of people. MMHP's announcement concerned the foundation for future investigation.
The most important limitation is the type of evidence. BioSpectrum Asia published a launch report, not a clinical-outcomes evaluation. It described an initiative and its early progress rather than testing a dietary intervention. The announcement should be assessed on those terms.
The reported 10,000-sample milestone was project information relayed by the publication. It was not independent verification of data quality. The report also did not provide sample demographics or recruitment details, so the size of the target does not establish how well the planned database would represent different populations.
The timetable needs its own boundary. The report does not establish that the one-million-sample goal was later completed. Looking back at the proposed schedule is therefore not the same as confirming its delivery. Neither a target nor the passage of its proposed schedule is proof of a completed resource.
Methods are another question. The launch account did not supply a study protocol, which leaves the reader without a detailed basis for assessing how the samples would be collected and compared. A useful evaluation would ask whether collection and analysis were consistent. It would also ask what lifestyle and medical information accompanied the samples.
Those questions matter when judging the meaning of scale. A very large dataset is a numerical description of a resource. Evaluating that resource also means asking what each record represents and whether comparisons are appropriate. Sample count is one consideration in that assessment, not a substitute for the rest.
The launch report did not set out consent or data-access terms either. Those are separate questions about how a proposed research database would operate. They should not be answered by assumption simply because the project involved several institutions.
Finally, the announcement did not validate individual diagnostic thresholds or demonstrate clinical benefits from using the database. The project leaders' comments about future medical applications were forward-looking views. That is the key boundary when interpreting a personal microbiome result: a population research resource and an individual clinical decision tool are not interchangeable.
The realistic dietary takeaway is modest: the 2019 announcement offers no basis for changing foods or supplements because of an individual microbiome result. Its value as news was the ambition to strengthen the research foundation. That ambition deserves attention without becoming another rule about what belongs on a plate.
People do not need to treat a large sequencing target as a reason to pursue an ideal microbial profile. The relevant question for any proposed change remains whether evidence supports that change and its intended outcome. When a test result is being used to consider restrictive eating or a change in care, a qualified healthcare professional can address the personal decision.
The lasting lesson from the plan is that better research resources and better everyday advice are connected goals, but they are not the same achievement. A larger map may support future answers. It should not be mistaken for those answers already being available.
A million-sample ambition can draw attention away from the separate question of whether a finding is ready for everyday use. DigestGenius addresses that gap with free educational articles and in-depth resources on digestion and microbiome science, helping readers distinguish promising research from established evidence when gut-health advice seems contradictory.
DigestGenius publishes research-led guidance on digestion, the gut microbiome, fiber, probiotics, gut-brain signaling, inflammation and everyday digestive wellbeing.



Explore clear, research-led guidance on digestion, the microbiome, food, fiber, probiotics and the gut-brain connection.
read the blog