[adegenet-forum] Maximum values for minimum distance neighbouring graph (network graph type 5)

[Jombart, Thibaut]

Hello again, 

indeed setting the default to zero would yield the same result. The interest of providing the minimum distance, however, is  to be able to define graphs by distance classes. This might be used to investigate spatial genetic patterns at different scales.

About the maximum value, I'm afraid there's no generic answer. What I've done in a data analysis of mammals was considering the distance at which the home-range of the individuals overlapped. For plants, this could be the dispersal distance of a seed. I think in any case a biologically-defined distance would be better in this case. If you find something suitable for your data, don't hesitate to share it on the forum ;-)

Best regards

Thibaut.


________________________________________
From: adegenet-forum-bounces at lists.r-forge.r-project.org [adegenet-forum-bounces at lists.r-forge.r-project.org] On Behalf Of Nevil Amos [nevil.amos at gmail.com]
Sent: 07 September 2010 09:38
To: adegenet-forum at lists.r-forge.r-project.org
Subject: [adegenet-forum] Maximum values for minimum distance neighbouring graph (network graph type 5)

  Jombart et al 2008 suggests the use of the minimum distance graph
where there is clustering of individuals used in an spca.

In this case the minimum value is set to the minimum distance between
individuals ( would setting it to 0 have the same result, I think so)

How does one choose the maximum value?  I have tried varying this value
and end up with quite different s.image plots of the resulting scores.

I had a look at Legendre's  book but this did not help me.

cheers

Nevil Amos


_______________________________________________
adegenet-forum mailing list
adegenet-forum at lists.r-forge.r-project.org
https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/adegenet-forum