I'm having some trouble describing what the programs MrBayes, PAUP*, and Garli are actually doing between the inputting of sequences and the tree output. I have a rudimentary knowledge of each of these programs, (Garli and Paup create multiple trees and compare the sequences to each of those trees and determines the best most likely/most parsimonious and MrBayes creates a single tree by looking at all the possible branching options and chooses the most probable one for each point) but after a certain point I get lost with the technical stuff. Can any one please offer me a simplified explanation of how each of these programs comes to the answers they give and maybe the benefits of one program over the other?
In that same vein, if my data set is 75+ for a gene and 200+ for another (all the same species) would it be worth looking into performing a neighbor joining test? Some people that I have talked to stated that there isn't enough information with the one species to perform it while I have see papers do it using one species. Any incite one way or the other would be much appreciated.