FIG 8.
Per base substitution rates. (A) Sabin type 1; (B) Sabin type 2; (C) Sabin type 3. Vertical bars, 95% HPD; red bars, known attenuating sites. When the rate inference is performed independently for each site, most sites are consistent with evolution at the neutral substitution rate (2.8 × 10−5 nt per site per day (9), as indicated by the 95% HPD. Sites under strong selection pressure show substitution rates that are significantly above the prior estimate for the neutral rate. Note that our estimate for position 2493 of Sabin type 3 is dramatically broadened to account for the possibility of nonnegligible variation in vaccine stock (42, 52). The results in panels A to C correspond to those in Tables 2 to 4, respectively.