Human Microbiome Project

The Human Microbiome Project was launched by the National Institutes of Health in 2007 with the mission to generate the resources and expertise needed to characterize the human microbiome and analyze its role in health and disease. The NIH approved a budget of $170 million for this project over five years, providing support for a number of centers and institutes around the United States, including Baylor College of Medicine. The HMP serves as a "road map" for discovering the roles these microorganisms play in human health, nutrition, immunity, and disease in diverse niches of the human body. A major goal of the HMP is the metagenomic characterization of microbial communities from 300 healthy individuals over time. The HMP is focused on studying the microbes residing in five body areas: skin, mouth, nose, colon and vagina.

The HMP is a coordinated effort being conducted at 80 institutes across the United States. Baylor College of Medicine is the only site where all aspects of the HMP, from human sampling to sequencing and data analysis, were performed. Initial efforts focused on technological issues involving the development of resources and procedures to accomplish the task of generating and analyzing vast amounts of data.

Because a goal of the HMP is to define a healthy human microbiome, a major challenge was the identification and selection of 300 “normal” subjects. The healthy adult volunteers that researchers recruited for this project were not obese, not on medications, and did not have any chronic health problems or diseases; even minor gum disease was enough to exclude a subject from the study. The clinicians collected multiple samples from different areas of the skin, mouth, nose, stool, and vagina of the volunteers (for a total of 15 sample sites from men and 18 from women) and sampled subjects up to three separate times over the course of a little over one year in an effort to access stability and diversity of the microbiome over time.

Over 11,000 human specimens were obtained. Scientists then purified and sequenced the DNA from them and used information from the bacterially-encoded 16S ribosomal RNA gene to identify and quantify the relative abundance of the bacteria in each sample. Subsequently, additional whole genome sequencing was performed on about 800 of the samples to learn about the genes that encode metabolic functions provided by the microbial communities residing at different body sites. To date, the HMP has generated 3.5 terabytes - or 3.5 trillion bytes - of data, or more than 1000 times the amount produced by the original Human Genome Project.

Researchers from the Department of Molecular Virology and Microbiology (MVM) have been instrumental in various aspects of the HMP. MVM faculty members Dr. Joseph Petrosino, Dr. Sarah Highlander, Dr. Wendy Keitel, and Dr. James Versalovic (who holds a primary appointment in the Department of Pathology) were involved with the HMP in the early phases of the project's design. They served on HMP Working Groups that established the criteria used to select "normal" subjects and determine the optimal number of body sites and subjects to sample, as well setting guidelines for ethical considerations in subject recruitment. They also worked on developing and testing protocols for standardized sample collection and processing.

Half of the subjects who participated in the HMP were recruited by Baylor. A large part of this success was due to Dr. Keitel's expertise in volunteer recruitment as a result of her work in directing the Vaccine Research Center at Baylor. She served as a principal investigator of the sampling aspect of the Baylor HMP.

The coordination of the human sampling efforts at BCM and at Washington University, in St. Louis, MO, was led by Dr. Versalovic, director of the Texas Children's Microbiome Center, who helped design the methods of clinical sampling. He also directs studies examining the role of the microbiome on healthy children, as well as on children with short bowel syndrome, pediatric irritable bowel syndrome, and recurrent abdominal pain.

Co-principal investigator of the BCM HMP, Dr. Highlander, developed mock communities of bacterial cells and bacterial DNAs in order to test and refine the sequencing and bioinformatics methods later used to characterize the human samples. She also was involved in generating nearly a quarter of the 800 bacterial reference genome sequences that formed a database to map sequences obtained from human samples. She serves as the BCM representative to the International Human Microbiome Consortium.

Dr. Joseph Petrosino, co-principal investigator of the project, is responsible for coordinating the metagenomic and microbiome research and development efforts across BCM and the Texas Medical Center. He serves as the director of the Alkek Center for Metagenomics and Microbiome Research (CMMR).

The work on the HMP has yielded fascinating and important insights into the human microbiome. In 2011, the HMP published a report in the journal Science describing sequencing of the first 178 of an anticipated 3000 microbial reference strains, which will serve as a resource for metagenomics studies. Recently, two major papers describing results from the first 242 healthy adults were published in the journal Nature, along with a number of additional publications in PLoS One and other journals. The reports indicate that there is a much greater diversity - both from person to person and between different sites within an individual - than previously realized.

A key finding is the extensive variation in microbiome composition from person to person, even in healthy individuals. There is not a single "normal" or core microbiome; everyone has a personalized microbiome. Nevertheless, different bacterial species may be doing similar jobs in different people. For example, the metabolic processes required to digest complex carbohydrates in the gut may be performed by different bacteria in different individuals.

Notably, researchers observed that most communities of microbes are distinct from one another (such as those on the skin, in the intestine, mouth, and vagina) and do not appear to mix, and not every body site contained members of all the major groups, or phyla, of bacteria known to colonize the human body. Rather, specific groups of microorganisms colonize distinct anatomical niches. Each body site showed a few core or "signature" bacteria with characteristic genes linked to that site, although the relative abundances of these bacteria varied from person to person.

The collections of microbes within different body regions show a surprising degree of diversity. Oral and stool samples had the highest numbers of different types of organisms, followed by the superficial skin samples. The vagina had the least bacterial diversity. There were also substantial differences in the diversity and composition of microbial communities between samples taken from different sites within the same body region, for example, from different areas of the skin.

When the researchers examined which microbes were present, they did not find genes commonly associated with highly pathogenic bacteria, but they did find organisms that are considered opportunistic - ones that can cause disease under certain circumstances. For example, they found Staphylococcus aureus in the noses of 30 percent of the subjects and Escherichia coli in the stools of 15 percent. The absence of disease-causing organisms from the microbiome suggests that people acquire these pathogens from other sources.