The characteristics of predicted site pairs for the estimated FDR≤10% are shown: positions of sites on the primary sequence of S-protein (site1 and site2), the upper p-values (p-value), values of the epistatic statistic (epistat), numbers of consecutive pairs of mutations (#consec. pairs of mutations), numbers of mutations in consecutive pairs of sites (#mut. in consec. pairs in site1 and site2), total numbers of mutations on the internal tree branches in sites (#mut. in site1 and site2), the expected value of epistatic statistics (exp. epistat) and the corresponding standard error (SE). The true predictions are highlighted in bold.