Assignment of Individuals to Different Populations

[laiaju...@gmail.com]

Hello everyone, I have been running some analyses, but I am realizing that depending on the filter for my individuals, the function gl.run.snmf assigns my individuals to different populations, and I would like to understand why. In the graph, you can see that in one instance it assigned the GOM individuals to the same cluster as COL, but in another, it did not. It makes sense that the GOM individuals are different from COL and MEX and the other locations because they are individuals from the Atlantic and not from the Pacific. However, I would like to understand what the assignment in the Q matrix depends on.

I am ussing the same code to run both gl objects:

yft_snmf1 <- gl.run.snmf(x=gl3, minK=2,
                 maxK=6, rep=10)

Q1 <- gl.plot.snmf(snmf_result=yft_snmf1, plot.K = 3, ind_name=F, color_clusters = viridis(K))
gl.map.snmf(gl3, qmat = Q)

Additionally, I am trying to use the function gl.run.popcluster, but I am getting different errors with the two distinct gl objects, and I don’t really understand what is happening.

I greatly appreciate your help.

Laia

[Jose Luis Mijangos]

Hi Laia,

Filtering is a really important step when analyzing data, but there’s no one-size-fits-all approach. Different filtering strategies can influence downstream analyses in different ways—I've attached some papers that explore this.

I’d recommend using PCA, since it’s a model-free approach, to help guide your decisions. Plotting the first three principal components in a 3D plot could give you a clearer picture of the data structure. Also, increasing the number of replicates to 20 would help strengthen the analysis.

Cheers,
Luis

[Laia Juliana Muñoz]

Thank you so much Jose Luis, I really appreciate all your help.

Best,

Laia

--
You received this message because you are subscribed to the Google Groups "dartR" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dartr+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/dartr/0a937a1d-137e-4b72-b913-932800371295n%40googlegroups.com.

[Bernd.Gruber]

Hi

As Luis said it depends on the fitering steps you used.

 

One thing that especially comes to mind is missing data (because individuals can look similar that have similar missing data at a number of loci.

 

So in your case it might make sense to look for patterns in your missing data,  between the groups of individuals.

 

Cheers, Bernd

To view this discussion visit https://groups.google.com/d/msgid/dartr/CAHxmJvLHOb%3DFOaa6avznCHA-kcJCwCLmg3j%3DZ9K2VeaphN3AYg%40mail.gmail.com.

[Laia Juliana Muñoz]

Thank you so much Bernd.

Best,

Laia

To view this discussion visit https://groups.google.com/d/msgid/dartr/SY4PR01MB79399292399A46353AA55172D4FA2%40SY4PR01MB7939.ausprd01.prod.outlook.com.