Unusual clustering results for Caucasian oak Quercus calcarea

[Денис Шуваев]

Hi, everyone!

There are several species of oak trees in the Caucasus.

Quercus robur

Q. calcarea

and three other species too (macranthera etc, it doesn't matter which ones specifically).

Quercus robur is well divided into geographical regions of the Caucasus on Structure.

But we do not see differentiation in the Structure of  Q. calcarea either in the variant of

runs of one species (file: "One_species_Q_calcarea_in_Structure_analysis.pdf") 

or in the variant of a run with  Q. robur (file: "Q.robur and Q. calcarea.pdf").

One might decide that there is simply no differentiation of Q. calcarea.

However, PCOA-analysis (gene-distances) showed a weak trend in west-east direction across the Caucasus for Q. calcarea.

It is even more interesting when we add the other three species of Caucasian oaks to Q. calcarea and pedunculate oak. 

Then this weak differentiation trend for Q. calcarea begins to show up in the Structure (file: "Q.calcarea_robur, and other oak species.pdf").

I don't understand why this is happening. Please explain to me the logic of this program. 

Why in a separate analysis of the Structure for Q. calcarea there is no differentiation? 

In a joint analysis with pedunculate oak there is no differentiation of Q. calcarea either.

But in the analysis with three other oak species this weak differentiation of Q. calcarea appears.

Let me make it clear right away that this is not an artifact of the analysis; any hybrid samples

were removed from all of these analyses.

Please help me understand this result from the point of view of the logic of the program's purpose.

 One_species_Q_calcarea_in_Structure_analysis.pdf

 Q.calcarea_robur, and other oak species.pdf

 Q.robur and Q. calcarea.pdf

--

Best regards, Dennis