COLD SPRING HARBOR, NY (GenomeWeb News) – Researchers are using human reads from metagenomic sequencing studies to begin looking at how host genetics influence microbial communities on and in the human body.
Ran Blekhman, a researcher with Cornell University and Weill Cornell Medical College, presented preliminary findings from this search during a session on the genetics of complex traits at the Biology of Genomes meeting here yesterday.
While environmental factors are expected to play a part in the types and abundance of bacteria found in and on the human body, he explained, host-microbe interactions carved out by human genes are thought to affect these communities as well.
Such interplay is particularly intriguing given that human microbiome composition seems to correspond to some chronic host conditions including inflammatory bowel disease, obesity, and diabetes.
Along with studies looking at the relationship between host genetics and disease and between microbial communities and disease, Blekhman said, researchers are now beginning to look more closely at the way that host genes relate to these corporal communities.
To look at this in more detail, Blekhman and colleagues from Cornell University and the Broad Institute took advantage of sequence data generated at 18 body sites each for 100 or so individuals through the Human Microbiome Project.
Samples from each of the body sites, organized into five so-called "super sites," had been tested using metagenomic shotgun sequencing to get an idea of which bacterial species are present and prevalent in the corresponding microbial communities, Blekhman explained.
But the researchers didn't initially have access to the sorts of participant genetic information needed to look at the host genes and pathways influencing each individual's microbiomes. They were able to get around this problem by taking advantage of the fact that some host DNA found in microbiome samples had been sequenced alongside bacterial DNA from these sites.
After doing quality control on these "human contamination" reads and mapping them back to the human reference genome, the team had enough sequence to cover the human genome to a depth of around 10 times per person — information that made it possible to find variants that could be used in a genome-wide association study related to microbiome features.
Using more than four million SNPs found in the human sequence data, investigators carried out an association analysis looking for sites in the genome with ties to 551 microbiome traits at 15 of the 18 body sites sampled.
Despite the relatively small sample size, the team was able to track down a few dozen signals showing genome-wide significant associations with microbiome traits including some loci previously implicated in studies of cholesterol and blood sugar levels, diabetes, Crohn's disease, or other complex diseases.
For instance, signals in or around type 2 diabetes-related gene PCSK2 appear to be associated with levels of bacteria from the Bacteroides genus in the gut, while an inflammatory and autoimmune-disease related gene called CXCL12 turned up amongst the genes linked to skin levels of bacteria from the Granulicatella genus, which can cause inflammation.
More generally, many of the genes that turned up in the team's analysis belonged to pathways involved in processes such as immune function, inflammation, and cell-cell interactions.
Together, such findings "highlight the complex molecular mechanisms involved in host-bacteria interactions, how these interactions may be affected by host genetic variation, and the possible impact of the microbiome on susceptibility to complex human disease," Blekhman and co-authors wrote in the abstract accompanying yesterday's presentation.