NEW YORK — By partnering with scientists and communities across the world, researchers can help boost representation within microbiome data repositories to ensure the benefits of such work are for all.
Two recent analyses have highlighted disparities in representation in microbiome studies. One found that microbiome data from European or North American individuals is overrepresented in genomic data repositories as compared to populations in Asia and Africa, while another showed that that overrepresentation affects how researchers interpret their findings.
Both studies highlight the need for more data representing a range of populations. That, researchers said, should be done by working with community members and colleagues in underrepresented regions. At the same time, such partnerships can guide researchers to certain microbiome research questions the community might want to study.
"When we're going to places and we're working with colleagues and communities where they don't have the types of research resources that we do, I think it's really important that we engage with community members who can help us shine a lens on the problems and issues that are important," Ami Bhatt, an associate professor at Stanford University, said.
In one of the recent analyses, a trio of researchers from the University of Minnesota scoured three genomic data repositories to find most of the nearly half a million microbiome samples came from Europe or North America, while other regions — particularly central and southern Asia — were underrepresented in the repositories.
For their analysis, the researchers teased out metadata on 444,829 human microbiome samples from 2,592 studies from the Sequence Read Archive, DNA Data Bank of Japan, and the European Nucleotide Archive. As they reported recently in PLOS Biology, they could trace country of origin for 86 percent of the samples.
Of these, 40.2 percent came from the US, even though it only makes up 4.3 percent of the world's population. Altogether, samples of European or US origin made up 71.2 percent of repository samples, while those regions comprise 14.3 percent of the global population.
At the same time, countries in Central and Southern Asia were underrepresented. Less than 2 percent of microbiome samples hailed from those regions, even as about a quarter of the world's population lives in those regions. Sub-Saharan Africa, meanwhile, was also underrepresented, accounting for 4.2 percent of samples, though the region includes 14 percent of the world's population.
Meanwhile, Bhatt and her colleagues found that by databases not including a range of global populations, studies of the microbiome may then miss key variations.
She and her colleagues examined the gut microbiomes of nearly 200 women living in urban and rural parts of South Africa and found that the microbiomes of both South African populations fall taxonomically between those observed in high-income populations and traditional populations, while still differing from one another, as they reported in Nature Communications.
Further, many of the microbes they uncovered within the South African women's gut microbiomes were not included in reference collections. This, they found, could influence how scientists interpret their results. Previous studies had suggested that alpha, or within-sample, diversity is higher among the microbiomes of people with traditional lifestyles, while beta, or between-sample, diversity is lower, as compared to the microbiomes of industrialized populations. But Bhatt and her colleagues instead suggested that the pattern could be due to a lack of representation of microbes more commonly found among populations from outside the industrialized regions that are represented in reference databases.
They compared the beta diversity of the South African and other gut microbiome samples using two measures, including a reference-agnostic approach. With a reference-based approach, the rural South African populations' microbiomes had lower beta diversity than a Swedish population, but the reverse was true in the reference-agnostic approach analysis, suggesting that difference may reflect the contents of the reference collection.
"Coming to the erroneous conclusion that beta diversity is smaller in quote-unquote non-Western populations is the consequence of not actually being able to categorize a lot of the sequence data that was generated in some of the first sequencing experiments," Bhatt said.
There has been an uptick in recent years in the number of microbiome samples from outside Europe and North America in genomic databases, the Minnesota researchers noted in their analysis. But more are needed, and they further called on stakeholders to prioritize and promote increased global representation of microbiome samples, such as through the Human Heredity and Health in Africa consortium.
Researchers from Europe and North America themselves could further do more to boost that representation by partnering with researchers and communities around the world.
Ilana Brito, an assistant professor at Cornell University who led a study examining mobile genes within the microbiomes of Fiji Islanders, said there is interest among many populations to learn more about the microbiome as well as in microbiome-based therapeutics, as they are often viewed as accessible treatments.
"I think there is excitement about understanding these diverse organisms that we don't typically see here, but are present in other populations," Brito said. "And so, partnering is actually quite easy, I think."
Bhatt, for instance, worked with other researchers — both direct scientific collaborators and other researchers from slightly different fields but working in the same region — as well as community advisers as part of her microbiome study in South Africa. Input from researchers and community members, Bhatt said, helped refine the design of the study.
Such engagement could also focus research on local concerns and local benefits. "I think it's not only ethically correct to do, but I actually think it also makes for much better research because at the end of the day, I think many of us would like our research to answer fundamental questions, but also solve problems that real people have," Bhatt said.
She added that she is also returning to South Africa to discuss the results with her partners as well as planning follow-up studies, including ones examining obesity and menopause in relation to the microbiome.
"I think with international collaboration, there's a chance of running the risk of it being extractive," Cornell's Brito added. "But if the partnership is done in a mutualistic, respectful, and non-extractive manner, I think it can work out really well."
She, for instance, has organized a symposium and workshop at Cornell that includes trainees from international sites, with the costs of attending highly subsidized. "It's an opportunity for everyone to come together and share expertise," Brito said, "and to motivate and to grow the interest in looking at the microbiome's role in health topics that are beyond those that are experienced in resource-rich areas."
Stanford's Bhatt added that a good piece of advice she received when she started in global health was that going to somebody else's country is like going to their home. "You don't walk in and tell them how you think you should do things or how you think they should do things," she said. "You ask, 'How can I help?'"