The range of aLRT results (approximate likelihood ratio test)

[Michael]

Dear all,

I tried to use both aLRT statistics (command line: -b -1) and SH-like (command line: -b -4) to get branch supports. However, I got a few questions about using aLRT.

1. Is SH-like also one kind of aLRT?

2. Is SH-like (-b -4) more commonly used (and more recommended) in evaluating branch supports than aLRT (-b -1)?

3. I know the range of SH-like branch supports is from 0 to 1, but what is the range of aLRT results (-b -1) and how should I interpret them? I ran aLRT (-b -1) for my data and I got the results ranging from 0 to 340.205, which I have no idea how to interpret. How much of an aLRT (-b -1) value can be considered a good one (such as corresponding to SH-like >0.7 or bootstrap >70)? 

4. In addition, I got a much larger number of branches with a high branch support (SH >0.7 or bootstrap >70) by using SH-like (-b -4) than using bootstrap (-b 100). Why did I get the inconsistent results and is there any way to resolve this inconsistency?

Any advice would be highly appreciated!

Best wishes,

Michael