Clade-specific varying effects

[ScottClaessens]

@ErikRingen We have talked briefly before about including clade-specific varying effects. I think you had experimented with this at some point. Could you perhaps dig out your Stan code implementation?

I think this would be a nice feature if we can implement it correctly.

[ScottClaessens]

@ErikRingen I started my new postdoc recently at the University of Bristol, where I'm hoping to use this package. I'm particularly interested in exploring the possibility of clade-specific varying effects. Do you have the old Stan model where you experimented with this? Were you able to get it to run?

[ErikRingen]

I couldn't find that code with an initial search, but I remember it was quite crude, just stratifying the tree and putting hierarchies on the coefficients. Would be happy to work together on getting a flexible implementation.

[ScottClaessens]

Thanks Erik. I had a think about how to implement this today. Looking at the current Stan code, I'm not sure if clade-specific varying effects are possible alongside the caching, since all of the computation is done in advance assuming that the parameters stay the same across the whole tree. But this isn't necessarily the case if we allow the parameters to vary in different clades.