In addition to the increasing focus on the human genome and human diseases at the AGBT 2013 meeting, there were a few very interesting microbiome presentations. Kjersti Aagaard, Baylor College of Medicine, gave a great presentation suggesting that the human placenta is not a sterile organ, as commonly believed. Her group found about 330 genus and 730 species in 48 placental samples after Illumina sequencing, producing about 100 million reads per sample, or about 4 terabases of data. MGRAST was used to map the data. Amazingly, none of the placental samples had any indications of infection, and sampling was rigorously controlled for contamination. So apparently, the human placenta harbors a non-pathogenic commensal microbiome which was the most similar to the oral cavity microbiome and not with the vaginal microbiome. When the species level similarity analysis was repeated by pathway analysis, the occurrence of pathways in the placental microbiome were more similar to the vaginal microbiome. She found an interesting correlation with the type of organisms present in the placenta with remote infections. An increased risk of pre-term birth and periodontal disease has been reported, so what is the link? Intercellular bacteria were present at the maternal-fetal interface, in the basal plates of the preterm and full-term gestations. Someone from the audience asked the obvious question, are fetuses sterile? Kjersti’s expectation was that the fetus is not sterile and this microbiome may have something to do with immune tolerance in neonates.
In the earlier part of her presentation on the vaginal microbiome, she had found a surprisingly accurate way for detecting pregnancy, by sequencing the vaginal microbiome! The microbial population change associated with pregnancy correlated very accurately with the onset of pregnancy.
Aargaard and Barbara Methé, from the JCVI gave an update on the Human Microbiome Project (HMP) that has the ambitious goal to map the incredible microbial diversity oin humans. To develop a functional annotation of the microbiome from 300 healthy adults, samples from 15-18 body locations such as the GI, viagina, oral and nasal cavities, skin and blood, as well as the sequencing of 3,000 reference strains have been completed. A human body harbors trillions of microbes, and we inherit about 2-4 M genes from these microbes. Culture-based methods pick up about 8-10% of these, therefore 16S rRNA or whole-genome sequencing becomes an essential tool to map the microbiome and understand how it varies by body niche, pregnancy, gender, diet, race, disease burden, among other conditions. The project has been highly successful, producing about 800 publications.
Barbara showed extensive data comparing microbial diversity with a series of plots that compared ubiquity (0-100% occurrences in samples) versus abundance (0.01% to < 100%, percentage presence in a sample). Whole-genome sequencing provided additional analysis options, including performing functional profile for KEGG orthologs. When doing species comparisons, a long tail is common (low frequency and high diversity) but with the KEGG analysis, the long tail becomes less evident. KEGG analysis is useful for seeing microbial differences such as antibiotic resistance, pathogenesis, and N-glycan biosynthesis.
More information on the tools used for microbial sequencing was provided by Jonas Korlach from Pacific Biosciences and Stephan Schuster, Penn State. Jonas discussed a new assembly approach that started with the longest reads delivered by PacBio (> 10 Kb) as a seed to which the other shorter reads (avg. 4 kb) are aligned to build a consensus. He showed examples with bacteria, plasmodium, and outbreak pathogens where this strategy works very well to produce contigs without the need for hybrid assembly strategies and without GC bias. Stephan Schuster talked about the sequencing environmental microbes. He stressed the importance of understanding complex microbial communities, for example, those that constitute biofilms. These are a form of “dark matter,” the largest biological matter on earth about which we have very little information—more than 99% of the earth's microbiome remains undefined. By sequencing wastewater sludge samples on the HiSeq 2500, he found that with 2 lanes of 400+ bp from overlapping reads, he approached saturation for identification of microbes in this complex community. Stephan used the MEGAN software for microbial identification.
Stephan presented an interesting analogy, the island of Singapore would grow by 10 cm every year if human waste was not processed! Without microbial activity urban life would be untenable. Maybe the next AGBT should be on the island of Singapore…