Hi Martin,

I think there are two things to consider with this question:

1) Mitochondrial markers, being maternally inherited, are going to be more linked than 3 markers selected from the nuclear markers. You could argue that this might justify the use of a concatenated data set for the PTP.

2) However, the substitution numbers that the PTP uses to delimit species will be biased towards more rapidly evolving markers in a concatenated data set. If for example COI has a faster substitution rate than the other genes in your data set, then the output of the PTP is going to reflect this gene rather than the others.

I suggest you run the PTP model on the 3 genes independently and compare the results between the markers and between a concatenated data set. If the number of "species" delimited is within the confidence limits of each other, and also that the identify and membership of each of your analyses do not pose hard incongruences, then you are justified in using the total evidence approach. If they are incongruent, you need to question why this is. It might be that processes such as selection make a concatenated approach unjustifiable.

Hope that helps,

Cuong Tang

Hello Cuong,

thanks for your advises. You're right that run the analysis seprately for three mtDNA gene is more wise. I did the analyses of them separately and the results are congruent. Can I ask you one more question? in PTP webserve, an downloadable output of annotated tree should show the tree with delimited species red marked, similarly to GMYC output tree (although as a picture there)? i am able to open it only in Treeview and there is nothing like that.

thanks in advance

Martin