[adegenet-forum] Very different numbers in groups with same paramters using find.clusters

[Thibaut Jombart]

Hello,

the K-means algorithm, used in find.clusters, uses random initial
conditions which may lead to converging to different solutions. This can be
the case if the data is noisy, and typically is the case when there are no
clusters.

The usual way to stabilise the results is increase the number of starting
points, which I think defaults to 10. Try setting n.start = 50 - that may
help.

Best
Thibaut


--
Dr Thibaut Jombart
Lecturer, Department of Infectious Disease Epidemiology, Imperial College
London
Head of RECON: repidemicsconsortium.org
sites.google.com/site/thibautjombart/
github.com/thibautjombart
Twitter: @TeebzR <http://twitter.com/TeebzR>
+44(0)20 7594 3658

On 11 January 2017 at 09:52, Markus Ruhsam <M.Ruhsam at rbge.ac.uk> wrote:

> Hello,
>
>
>
> I want to do a DAPC on a dataset containing 1548 SNPs starting with the
> find.clusters command but get wildly different numbers in each group when I
> repeat the command using exactly the same parameters. Does anybody have an
> explanation for this?
>
>
>
> Thank you
>
>
>
> Markus
>
>
>
> > grp <- find.clusters(rul, max.n.clust=10)
>
> Choose the number PCs to retain (>=1):
>
> 400
>
> Choose the number of clusters (>=2:
>
> 3
>
> > table(pop(rul), grp$grp)
>
>
>
>                1   2   3
>
>   native       2 131  52
>
>   non-native  67   3   1
>
>   hispanica    5   0   0
>
>   odd         12 113  41
>
>
>
> > grp <- find.clusters(rul, max.n.clust=10)
>
> Choose the number PCs to retain (>=1):
>
> 400
>
> Choose the number of clusters (>=2:
>
> 3
>
> > table(pop(rul), grp$grp)
>
>
>
>                1   2   3
>
>   native       1   1 183
>
>   non-native  15  51   5
>
>   hispanica    0   5   0
>
>   odd          0  12 154
>
>
>
> > grp <- find.clusters(rul, max.n.clust=10)
>
> Choose the number PCs to retain (>=1):
>
> 400
>
> Choose the number of clusters (>=2:
>
> 3
>
> > table(pop(rul), grp$grp)
>
>
>
>                1   2   3
>
>   native      49 134   2
>
>   non-native   1   3  67
>
>   hispanica    0   0   5
>
>   odd         41 113  12
>
>
>
>
>
>
>
>
>
>
>
>
>
>
>
> Dr Markus Ruhsam
>
> Molecular Plant Ecologist
>
> Royal Botanic Garden Edinburgh
>
> 20A Inverleith Row
>
> Edinburgh
>
> EH3 5LR
>
> United Kingdom
>
>
>
> Tel: +44 (0) 131 248 2859 <+44%20131%20248%202859>
> Fax: +44 (0) 131 248 2901 <+44%20131%20248%202901>
>
>
>
> http://www.rbge.org.uk/science/genetics-and-conservation/markus-ruhsam-
> homepage
>
>
>
> The mission of the Royal Botanic Garden Edinburgh is to “*To explore,
> conserve and explain the world of plants for a better future*”
>
>
>
>
>
> [image: cid:image002.png at 01D09F78.9DFBA130]
>
>
>
>
> The Royal Botanic Garden Edinburgh <http://www.rbge.org.uk/> is a charity
> registered in Scotland (No SC007983)
>
>
> _______________________________________________
> 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
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.r-forge.r-project.org/pipermail/adegenet-forum/attachments/20170116/fd962797/attachment-0001.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: image001.png
Type: image/png
Size: 31595 bytes
Desc: not available
URL: <http://lists.r-forge.r-project.org/pipermail/adegenet-forum/attachments/20170116/fd962797/attachment-0001.png>