Table 2

Timing estimates when a variable rate of evolution was allowed. The estimates of the timing of the M group origin were consistent over a wide range of priors. The results are presented as the estimated year and the 95% CI. The timing of the root of the tree requires a prior, and if we changed this from a prior of 1930 ± 35 years, to 1910 or 1950 ± 35 years (shifting our prior by 20 years), our posterior distribution only shifted by a year or two. The prior on ν influences the extent of rate variability between parent and child branches; increasing or decreasing ν by a factor of 3 also had little effect on the outcome.

Predicted M groupU.S. B clade nodeM group priorν prior
Env gp160 tree
1930 (1911–45)1954 (1941–63)1930 ± 350.03
1930 (1911–44)1955 (1942–64)1930 ± 350.1
1929 (1907–43)1954 (1940–63)1930 ± 350.008
1928 (1906–43)1954 (1940–63)1910 ± 350.03
1931 (1911–45)1955 (1942–64)1950 ± 350.03
Gag tree
1927 (1888–51)1959 (1937–71)1930 ± 350.03
1929 (1889–52)1959 (1938–72)1930 ± 350.1
1922 (1989–52)1956 (1933–70)1930 ± 350.008
1922 (1880–48)1956 (1933–70)1910 ± 350.03
1929 (1890–52)1960 (1939–72)1950 ± 350.03