Hi Emma

There is a lot of work done in Quercus phylogenetics (look for Paul S Manos work for a world-wide view, or Petit et al. 2002a,b for the European oaks phylogeography. Could you be more specific about the species (clades) and why they are not correctly represented?  Where, and why, did you get the idea that ITS2 should work for you?

Cheers

Pablo

Hi Emma,

    It appears several issues have been mixed up in your question. First, phylogenetic tree is just a hypothesis. With only one data (used to propose the hypothesis) we can not say if a tree is correct or not. Second, several factors can affect the results of phylogenetic inference. Getting different results from ML and NJ (distance-based) methods  (I guess you used different programs) may be due to 1). you have used different DNA substitution models in ML and NJ analyses; 2). the program you used applied different strategies to treat alignments gaps and missing data. Do you have alignment gaps? 3). Event you used the exactly same model, it is natural to have different results from ML and NJ analyses. NJ compressed information in DNA sequence into a simple distance matrix, while ML try to maximize the likelihood of observing  the alignment and model parameters given the trees. NJ is statistically inferior.   4). To my knowledge, the current version of RAxML always have the parameter GAMMA enforced  when performing tree search. The GAMMA parameter that takes into account of difference of substitution rates among sites  may have large effect on phylogenetic inference. There are many more issues that can result in different tree from NJ and ML analyses. Following review may help.

Yang Z., Rannala B. 2012. Molecular phylogenetics: principles and practice. Nature Review Genetics 13: 303-314.

best