deep time tree branch lengths #265
|
Hello, I have a tree encompassing red algae, green algae and land plants, which has a deep evolutionary time with a root age of ~1400 MYA. I try to estimate lambda and alpha values with K = 4. If I use branch lengths in a 1 MY scale, CAFE5 immediately stops with the message "Failed to initialize any reasonable values". This is also the case when I only focus on a subclade, e.g. green algae. However, when I scale the branch length to 0.001 MY, it does converge (with some failed families) to a lambda value and independent runs more or less lead to the same lambda value. I wonder now whether it is ok to scale branch lengths like that, or whether this potentially violates an assumption of CAFE5. |
Replies: 1 comment
|
Mathematically it's fine, but don't forget that lambdas are generally stated in terms of millions of years. So if you divided all of your branch lengths by 1000, you should multiply the resulting lambda value by 1000 - and discuss with your adviser whether or not that value makes sense in an evolutionary perspective. |
Mathematically it's fine, but don't forget that lambdas are generally stated in terms of millions of years. So if you divided all of your branch lengths by 1000, you should multiply the resulting lambda value by 1000 - and discuss with your adviser whether or not that value makes sense in an evolutionary perspective.