There are some possible factors why it may happen:

- different rates of genes variability

- possible presence of different gene copies, which are not always the same

- some of mt-genes can be transferred to the nuclear genome

At first,  this phonemon is the common, as the above said, the different DNA have different Substituion rate, Besides, the confience (or bootstrap value) of different clade is whether is enough large to robust this  clade, different largrithm result from the same gene is whether is consisit? all will produce inconsitent result.

Dear colleagues：

Thank you for your help and advises. And sorry for not give the details. Here is some explanation about your concerns:

1.  What methods I used?

I used BI and ML to reconstruct the phylogenetic relationship, the figure below show the details of the results.

2.  Did numts (mtDNA transferred to the nuclear genome) or mutation rate variability contribute to the conflicts?

Probably not for nuts, we have tested this by check sequence. For substitution rate, yes, but this can’t alone be used to explain our problems.

3.  Problem of Dloop?

Dloop is more variable than coding region of mitochondrial genome, and sometimes align may be a problem especially to CSB domain (HV II), as it often contain repeat sequences and long indels, but we use ETAS domain (HVI), it is relatively more conserved and easier to align.

4.  Problem of concatenated cytb-Dloop?

Yes, I have noticed that a lot phylogenetic studies used concatenated cytb-Dloop as dataset to inference the phylogenetic relationship. It is not appropriate to use the same model for the coding and non-coding region especially when the phylogenetic results of the two region are not the same. But in my case, the basal part (clades close to root) are consistent across the two region, but Dloop couldn’t resolve the deep relationship considering the low bootstrap values. 

 I saw your results, at first the gap of your aligment is whether large or much, if so, the aligment is not fit  to reconstructe the phylogenetic tree.

Then, I suggest you concatenate the ND1 and Cytb, because on one hand, both code protein gene, on the other hand, the D-loop high variety.

 Finaly, the eventual results is familar with the result, you need check the mtDNA reconstruction.  and BI tree was constructed by MrBays and so on ,ML  by PhyML and RAxML.

The picture of your tree is beantiful , would you tell me the sftware to modify the tree fille and generate the pdf file.

Dear Artur and Liu,

Thank you for your suggestion, I am happy to discuss some problems I meet in my project as a rookie, again thank you for your helpfulness and kindness. Blew are some explanations and answers to your questions.

1.     Where are the conflicts?

  For cyt b, AB vs C, but for concatenated A vs BC. Network analysis suggest the same pattern

As you suggested, the other discordance (I mean the sequencing of A.B.C.D.E in the phylogenetic tree, the divergence order. And, A.B.C.D.E are geographically divergent and have relatively complex demographically history) among different markers could be attributed to insufficient phylogenetic signals and low BPP or bootstrap. And, my project mixed population and species level polymorphism. I will try IMa, adegenet, and DYIabc, thank you again!!!

2.     Maker I used and picture I made?

Cyt b 1070bp, Dloop 396bp, ND1, 965bp

The topology was modified from trees loading by FigTree v1.4.2 (http://tree.bio.ed.ac.uk/software/figtree/), the other work is done by Adobe Illustrator CS6. These tools are cool!!!

You could run your tree and alignment through PhyDesign (web app) to better understand what the rates of evolution of markers imply about their Informativeness at different depths of your tree.

Dear Artur, 

So far, I have to say that you taught me a lot and gave me much inspiration, a rule come to that no support no discussion foundation. Thank you for your valuable time. It meant a lot to a novice like me. Blow is my network results, and I combined the result of phylogenetic reconstruction and network, also demographical history, I had some concludion, thank you again.

Hi Jeffrey, thank you for your advice, I am trying to do the analysis.

By the way, if I have some final conclusion about the whole pattern, the result will be posted on the website.

Thank you very much all.

I really appreciated your efforts, thank you all for your time and energy!!