PCR Detection of Mimivirus

T article by Zhang et al. (1) reporting the failure to detect mimivirus (Acanthamoeba polyphaga mimivirus) in patients with acute respiratory symptoms was flawed in 2 ways. First, a single mimivirus was detected by PCR but was dismissed by the authors as colonization. Second, and of greater concern, the PCR used by the authors would not detect most known mimiviruses because of primer/probe mismatches. In fact, PCRs used in most mimivirus prevalence studies would fail to detect most mimiviruses, given the extensive genome sequence diversity of these viruses (2). We evaluated the PCR primers/probes used by Zhang et al. (1) for sequence similarity with 51 mimivirus genomes available in the public domain since March 2016, the date on which the article by Zhang et al. was submitted (3). Our in silico analysis found that none of the 25 genomes available from lineage A and 7 genomes from lineage B were complementary with their primers and probes, and only 5 of 19 lineage C genomes showed sufficient sequence similarity to permit successful amplification (Table). Overall, the PCR used by these authors would be predicted to detect only ≈26% of lineage C mimiviruses and <10% of published mimivirus genomes overall. As proposed by Ngounga et al. (4), the great variability across the 3 lineages of mimivirus genomes requires the design of lineage-specific primers to improve PCR sensitivity (4). Until PCRs are improved, mimiviruses might be best detected by metagenomic approaches that have recently seen success in terms of mimivirus detection. For example, mimivirus-like reads have been identified by metagenomics in 1 (8%) of 12 fecal samples from children with acute diarrhea (5); 5 (27%) of 18 prostatic secretion samples from persons with prostatitis (6); all DNA libraries tested from the plasma of patients with hepatitis (7); 0.0018% of total viral reads from nasopharyngeal samples from 210 patients with respiratory tract symptoms (8); and all tested fecal, mid-vaginal, buccal mucosal, and retroauricular crease samples, with a predominance of mimiviruses in fecal samples reaching ≈33% of the total viral reads (9). In metagenomics approaches, the number of reads detected and the distribution along the whole viral genome