Diagnostic metagenomics and its associated trail of publications are spreading across the world. Multiple clinical labs in the United States, Europe, and Asia have gone to considerable lengths to optimize and validate a range of protocols for agnostically detecting viral, bacterial, fungal, and eukaryotic parasite nucleic acid across a range of patient specimens to aid in diagnosis for particularly recalcitrant cases. Others see a role for diagnostic metagenomics as a frontline diagnostic to replace other microbiological testing. Areas covered: There are considerable barriers to adoption for diagnostic metagenomics, including analytical sensitivity, interpretation, actionability, turnaround time, antimicrobial susceptibility, clinical utility, laboratory workflow, trial comparators, cost, and reimbursement. Expert commentary: Metagenomics is unlikely to become 'one test to rule them all' any time soon, not least because it is not indicative of historical infection like some of the highest volume tests in the clinical virology lab, viral serologies. The high cost and low marginal utility compared to 'standard of care' diagnostics have forced metagenomics to be mostly used for last-ditch cases. However, waiting for such patients to declare themselves as being diagnostically challenging in turn likely lessens the diagnostic yield and actionability of the information. Significant reductions in the cost of metagenomic sequencing are required for it to move up in the diagnostic pipeline. This review covers these associated obstacles of metagenomics, arguing for a parsimonious role in last-ditch diagnostics and awaiting the answer of many outstanding questions regarding its adoption.
Keywords: Luddite; Metagenomics; clinical metagenomics; clinical microbiology; mNGS; next-generation sequencing; utilization management.