PCA results

[Laia Juliana Muñoz]

Hi all

I am trying to do my PCA, I am using the code: dapc1 <- dapc.genlight(gl11)

However, I dont understand why I can´t see all my individuals (192 from 3 different populations), I just have 1 sample in each population in the scatter. 

Somebody could help me please. 

Best,

Laia

[Zhian Kamvar]

Forgot to send this to the list *shakes fist at google*

---------- Forwarded message ---------
From: Zhian Kamvar <zka...@gmail.com>
Date: Thu, May 5, 2022 at 16:30
Subject: Re: [poppr] PCA results
To: Laia Juliana Muñoz <laiaju...@gmail.com>

Hello, 

This is likely because you chose the maximin number of Principle Components to use for DAPC. When that happens, your model will overfit your data and the discriminant axes will put such a vast distance between the groups that they appear to cluster into single points. 

During the step where you choose the number of PC for the DA, the rule of thumb is to choose the number of PCs that makes up ~80% of the variation. 

Hope that helps. 

Zhian

--
You received this message because you are subscribed to the Google Groups "poppr" group.
To unsubscribe from this group and stop receiving emails from it, send an email to poppr+un...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/poppr/1077e185-0df3-4d82-ad9d-e0953f56b700n%40googlegroups.com.