[adegenet-forum] Different population clusters with DAPC and STRUCTURE

Jombart, Thibaut t.jombart at imperial.ac.uk
Fri Oct 9 12:41:09 CEST 2015

Hi there,

differences in results could be explained by large differences in within group diversity, in which case the K-means may struggle to identify the 'right' clusters. One way to check this would be compute e.g. Hs in your different STRUCTURE clusters.

Alternatively, your data might deviate from the model assumptions made by STRUCTURE and lead to sub-optimal cluster definition in the model-based approach.

It would be useful to look at the data from a different angle: run them through a basic PCA to have an idea of the overall diversity, and also look at hierarchical clustering based on the Euclidean distances between individuals, e.g.:

h1 = hclust(dist(tab(x, freq=TRUE, NA.method="mean"))^2)

etc. (check the 'method' argument in hclust)

As a side note, Fst computed on clusters defined by any reasonable clustering method will always be significant (the testing is circular).


From: adegenet-forum-bounces at lists.r-forge.r-project.org [adegenet-forum-bounces at lists.r-forge.r-project.org] on behalf of Simon Dures [Simon.Dures at ioz.ac.uk]
Sent: 08 October 2015 17:00
To: adegenet-forum at lists.r-forge.r-project.org
Subject: [adegenet-forum] Different population clusters with DAPC and STRUCTURE

Dear Dr Jombart and others,

I am using DAPC, STRUCTURE and TESS to look at genetic structure across my population (lions), using 20 microsat markers.
The STRUCTURE and TESS results suggest k=2 but with substructuring of k=4. These results make very good ecological sense and have significant Fst.
When I run find.clusters the suggestion is for either 3 or 4 clusters. DAPC using k=4 gives a very nice output that also makes ecological sense, however the cluster membership is very different to those identified by STRUCTURE analysis.
DAPC appears to pick up more detail about the substructure in one instance, however it completely removes one cluster identified by STRUCTURE.
Running many different combinations of my data in structure with different settings never removes this additional group, and likewise in DAPC it never appears.
Do you have any suggestions as to what may be causing this and how I can resolve the conflicting results?
My thought are it may be either because DAPC is not based on genetic models and so is missing some details, or that DAPC is reducing some of the data too much, however I am not knowledgeable enough to really have a strong ideas about the theory involved.

I hope you are able to suggest some possibilities.
Many thanks for any help that you can provide

Kind regards

The Zoological Society of London is incorporated by Royal Charter
Principal Office England. Company Number RC000749
Registered address:
Regent's Park, London, England NW1 4RY
Registered Charity in England and Wales no. 208728

This e-mail has been sent in confidence to the named addressee(s).
If you are not the intended recipient, you must not disclose or distribute
it in any form, and you are asked to contact the sender immediately.
Views or opinions expressed in this communication may not be those
of The Zoological Society of London and, therefore, The Zoological
Society of London does not accept legal responsibility for the contents
of this message. The recipient(s) must be aware that e-mail is not a
secure communication medium and that the contents of this mail may
have been altered by a third party in transit.
If you have any issues regarding this mail please contact:
administrator at zsl.org<mailto:administrator at zsl.org>.

This message has been scanned for viruses by MailControl<http://www.mailcontrol.com/>, a service from BlackSpider Technologies<http://www.blackspider.com/>.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/adegenet-forum/attachments/20151009/e69127dd/attachment.html>

More information about the adegenet-forum mailing list