Statistical Design and Analysis of PCR Tests for Fast Mutating Viruses

Han, Y.; Sun, Y.; Hsu, J. C.; House, T.; Gent, N.; Hall, I.

Mutations in the SARS-CoV-2 virus has given rise to concerns how diagnostic tests, treatments, and vaccines are affected. This article shows mutations in the Spike gene (S gene), which is prone to mutation, can be differentiated using standard TaqPath PCR (polymerase chain reaction) test. The methodology proposed in this article can be used as an alternative tool for detecting S gene mutations when sequencing is not available. Overcoming patient-to-patient variability by conditioning on an interval control, our statistical methodology classifies each patient as infected by wildtype or by some mutant strain(s) with reasonable accuracy. Besides adding a new tool for tracking emerging mutations epidemiologically, being able to make such patient level calls may become important for treatment purpose, as there is evidence that some antibody treatments are less active in neutralizing the SARS-CoV-2 virus against certain mutant strains. Our algorithm can be continuously retrained as the SARS-Cov2 virus evolves.