Dating divergent sequences/tips

[ekankaka]

For HIV infection, diversity generally increases with time. But based on the viral load trajectory in natural HIV infection, early viremic infection also seems to have a lot of evolution (which reduces and increases again in late chronic infection). If true, divergent sequences with censored/missing dates would have equal chance of mapping to early or late viremic infection. How does bayroot handle this?

[ArtPoon]

The short answer is that bayroot does not handle variation in the rate of diversification at all.

You're making the assumption that rate of diversification is roughly proportional to viral load. This is not necessarily a bad assumption but it is complicated. For example, neutral evolution in terms of substitution (divergence, not diversification) should be independent of population size (mutation input is $N \times \mu$, probability of fixation is $1/N$, so substitution rate becomes $u$).

It is conceivable that there is a burst of diversification at an early stage of infection associated with viremia. However, this is exponential growth from a fairly homogeneous population. Have we ever seen a surge of diversity associated with viremia in HIV RNA? Off the top of my head, I can't think of an example of this, although I assume these samples would be quite difficult to obtain.