Parallel evaluation of broad virus detection methods

The testing for adventitious viruses is of critical importance during development and production of biological products. The recent emergence and ongoing development of broad virus detection methods calls for an evaluation of whether these methods can appropriately be implemented into current adventitious agent testing procedures. To assess the suitability of several broad virus detection methods, a comparative experimental study was conducted: four virus preparations, which were spiked at two different concentrations each into two different cell culture media, were sent to four investigators in a blinded fashion for analysis with broad virus detection methods such as polymerase chain reaction-electrospray ionization mass spectrometry (PCR-ESI/MS), microarray, and two approaches utilizing massively parallel sequencing. The results that were reported by the investigators revealed that all methods were able to identify the majority of samples correctly (mean 83%), with a surprisingly narrow range among the methods, that is, between 72% (PCR-ESI/MS) and 95% (microarray). In addition to the correct results, a variety of unexpected assignments were reported for a minority of samples, again with little variation regarding the methods used (range 20-45%), while false negatives were reported for 0-25% of the samples. Regarding assay sensitivity, the viruses were detected by all methods included in this study at concentrations of about 4-5 log10 quantitative PCR copies/mL, and probably with higher sensitivity in some cases. In summary, the broad virus detection methods investigated were shown to be suitable even for detection of relatively low virus concentrations. However, there is also some potential for the production of false-positive as well as false-negative assignments, which indicates the requirement for further improvements before these methods can be considered for routine use.