Privacy is a concern in human microbiome research. Identifying an individual using microbial DNA present in a sample that they provide is predicated on the relative stability of any given microbiome, its uniqueness to an individual, and the nature of the residual microorganisms that we leave in the environment. A growing body of evidence suggests that these three features of microbiomes could be exploited for forensic applications^1,2,3. However, any human sample collected for microbiome analysis contains more than just microorganisms, because we shed human cells constantly, and human DNA sequences are often present in microbiome datasets. To allay privacy concerns in the National Institutes of Health’s Human Microbiome Project, removal methods for human DNA reads were pioneered⁴. These methods use computational screening of sequences against a human reference genome to identify and remove them. Although such methods are in use today, we need a comprehensive database to capture the unique variation that exists among human populations to be certain that all human DNA sequences have been removed. Such a database does not exist. Now writing in Nature Microbiology, Okada and colleagues⁵ report that there is substantial potential for identification of participants using metagenomes sequenced from human faecal samples.