I have a reasonably high amount of mitochondrial sequences (~500) to align to each other. Has anyone ever had experience with evaluating which software is best for mitochondrial sequences ? MUSCLE, MAFFT, PRANK or something else ?

I am asking because I am trying to evaluate which is best. What I am doing is loading the resulting multiple sequence alignment in an R script, computing a neighbor joining tree (ML is not computationally feasible for this amount of seqs). I tried using ape or phangorm. Is there a way to compute the likelihood of the data given a tree?