A method to identify and quantify microbes according to their sequence data is reported online this week in Nature Methods. The approach reported in this study generated an accurate species-level profile of the human gut and will aid in our understanding of microbial communities and their role in health and disease.

To survey the diversity of microbes in an environmental sample, researchers often use variations in a single conserved gene sequenced from each microbe’s DNA as ‘fingerprints’. Microbes in the sample can be identified by grouping the fingerprint sequences by degree of similarity and comparing them with those of known species.

Peer Bork and colleagues extend this idea by using ten genes that are conserved to varying extents, and they show that these give far higher accuracy than the single-gene approach when assigning sequences to microbial species. The team’s freely available software automatically extracts the ten gene sequences from genomic data derived from random DNA fragments in mixed sample and performs species assignment. The method does not rely on previously sequenced genomes and can be used to profile unknown species. In the human gut, the researchers show that the unknown fraction of microbes makes up over half of the total diversity. By profiling these species using existing data, they find new species that are preferentially present or absent in patients with ulcerative colitis relative to unaffected individuals.