Note that their "slow molecular clock" estimate gives a most recent common ancestor around the end of December. They reject this because it "finds a root which

Note that their "slow molecular clock" estimate gives a most recent common ancestor around the end of December. They reject this because it "finds a root which seems very early given the data". It's also, of course, possible that this particular cluster died out entirely, got genetically bottlenecked, or only ended up causing a small proportion of cases just through random chance.