HI! I am using phylip to generate a big protein tree (more than 200 proteins) using a default setting, but the the tee showed very shot internal branches. Can anyone give a suggestion for me to improve the tree construction. I think phyml could be better but it is really take time. the tree looks good with treefast, but I think it can not meet the publish quality. Thanks!
If there are short internal branches, this is likely more of an issue with your data than with the method itself. It may be the case that you cannot resolve the relationships with confidence with your dataset. Have you estimated support values? Additionally, if the proteins are vary conserved (hence, not many phylogenetically-informative sites) or very have conflicting topologies, it could be an issue.
Also, what's the 'default' setting? Are you using corrected genetic distances? You might consider performing model selection on your data and correcting distances as opposed to using uncorrected ones.