Are there any methods optimized for Neighbor Joining and Maximum Likelihood phylogenetic tree reconstruction?
In our benchmark which was done with protein simulated data, we have seen that the strictplus automated method works better for tree reconstruction using Neighbor Joining while the automated1 heuristic method works better for the tree reconstruction using Maximum Likelihood. However, the user should keep in mind that this has not been tested in real biological data nor using DNA/RNA data. trimAl has other methods -gappyout and strict- that may be better in other situations.