Understanding best$eventData

[MSerrano]

Hi all,

I have got results for a 600sp phylogeny that look like the attached file, with 6 shift events. 
When I investigate the actual nodes and values of these shifts (different models) I get :

    node     time      lam1        lam2               mu1     mu2 index
1  596  0.00000 0.1067686 -0.02013810 0.02105032   0     1
2  679 20.98388 0.4043447 -0.04162876 0.01096694   0     2
3  615 31.33035 0.3531746 -0.03337251 0.03204575   0     3
4 1018 46.13333 1.4491488 -0.23970019 0.03578583   0     4
5  845 49.89349 0.7922181 -0.10910905 0.11037264   0     5
6  971 50.28458 0.8471259 -0.17031247 0.09767037   0     6

I'll really appreciate if we could discuss the following :

- Are node times reversed ? my tree goes from 56 Mya to 0 in present, and node 596 is the root with a time of 0.0, and the most recent shift occurred at node 6, which will correspond to 56-50.28 = 5.72 mya?

- What do lam1 and lam2 correspond to ? If lamb1 is the "previous" lambda value for the process and lam2 the actual change, I don't understand why my values are negative (-0.02...) if according to the plot diversification rates increase on these nodes. Or is the combination of lam - mu what makes the difference, but it is not evident from the table.

Thanks in advance,
MSerrano

[Dan Rabosky]

Hi-

You could think of node times as being reversed, though obviously it depends on your frame of reference. BAMM starts with time t = 0 at the root of the tree. You are correct in interpreting the shift time at your node 6.

lam1 and lam2 are parameters of the exponential change model. The model is: lam(t) = lam1 * exp(lam2 * t)

where lam(t) is the speciation rate as a function of time. So yes, lam2 can be negative - in fact, it will always be negative if a given shift leads to a dynamic with temporally declining speciation.

~Dan Rabosky

[Jamie T]

So is an upshift defined by lam2 being positive, and downshift defined as lam2 being negative?

[pasca...@gmail.com]

No, lam1 and lam2 are the parameters of the exponential change model for that rate regime. See Dan's response above. It is not in relation to the rate shift regime that precedes it.