fastStructure results are not looking good

[Vinod Kumar]

I have run faststructure for a dataset, the same I've used for PCA. In PCA, I can locate 4 different populations easily but when I run faststructure for k=1 to 10 and run chooseK script, the results are awkward, I am pasting the results here,

Script:
python structure.py -K 1-6 --input=finalpruned1 --output=genotypes_finalpruned1cv --cv=5 --full --seed=100

Script:
python chooseK.py --input=genotypes_finalpruned1cv

Model complexity that maximizes marginal likelihood = 6
Model components used to explain structure in data = 1
..

Marginal Likelihood = -0.8159680202 (k=1)
Marginal Likelihood = -0.6996965619 (k=2)
Marginal Likelihood = -0.6399267603 (k=3)
Marginal Likelihood = -0.6181590019 (k=4)
Marginal Likelihood = -0.6065174658 (k=5)
Marginal Likelihood = -0.5901150086 (k=6)

I think from MLs, I can see differences are greater up to k=3, but chooseK is resulting only k=1, I am not understanding what to do? Can anyone have a look on this?

Thanks,

[Vikram Chhatre]

chooseK is indicating that the optimal K is somewhere between 1 and 6, not 1.  It is often helpful to examine bar plots at each K to see what makes most sense given the biology of the species and history of source populations.

V

--
You received this message because you are subscribed to the Google Groups "structure-software" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structure-softw...@googlegroups.com.
To post to this group, send email to structure...@googlegroups.com.
Visit this group at http://groups.google.com/group/structure-software.
For more options, visit https://groups.google.com/d/optout.

[Vinod Kumar]

Thanks Vikram Chhatre. Thats great, I was thinking the final k is 1. Can you also suggest me that how to plot individual-wise admixture bar-plot like result of GUI version of Structure2.3.4 (Multiple line).

Thanks,

Vinod Kumar

--

PostDoc Fellow,

International Wheat Genome Sequencing Consortium (IWGSC),

NRC on Plant Biotechnology,

Indian Agricultural Research Institute,

New Delhi-110012

[Vinod Kumar]

Hi Vikram,
I've run data till k=14 and choosek is behaving like I've posted earlier, I don't know how long I am to go?
Do I have some problem in my dataset?

Model complexity that maximizes marginal likelihood = 14

Model components used to explain structure in data = 1

Thanks,

Vinod Kumar

On Sunday, September 20, 2015 at 7:52:50 PM UTC+5:30, Vikram Chhatre wrote:

chooseK is indicating that the optimal K is somewhere between 1 and 6, not 1.  It is often helpful to examine bar plots at each K to see what makes most sense given the biology of the species and history of source populations.

V

On Sun, Sep 20, 2015 at 10:11 AM, Vinod Kumar <kumar....@gmail.com> wrote:

I have run faststructure for a dataset, the same I've used for PCA. In PCA, I can locate 4 different populations easily but when I run faststructure for k=1 to 10 and run chooseK script, the results are awkward, I am pasting the results here,

Script:
python structure.py -K 1-6 --input=finalpruned1 --output=genotypes_finalpruned1cv --cv=5 --full --seed=100

Script:
python chooseK.py --input=genotypes_finalpruned1cv

Model complexity that maximizes marginal likelihood = 6
Model components used to explain structure in data = 1
..

Marginal Likelihood = -0.8159680202 (k=1)
Marginal Likelihood = -0.6996965619 (k=2)
Marginal Likelihood = -0.6399267603 (k=3)
Marginal Likelihood = -0.6181590019 (k=4)
Marginal Likelihood = -0.6065174658 (k=5)
Marginal Likelihood = -0.5901150086 (k=6)

I think from MLs, I can see differences are greater up to k=3, but chooseK is resulting only k=1, I am not understanding what to do? Can anyone have a look on this?

Thanks,

--
You received this message because you are subscribed to the Google Groups "structure-software" group.

To unsubscribe from this group and stop receiving emails from it, send an email to structure-software+unsub...@googlegroups.com.

[Vikram Chhatre]

What do your bar plots look like?

To unsubscribe from this group and stop receiving emails from it, send an email to structure-softw...@googlegroups.com.

To post to this group, send email to structure...@googlegroups.com.
Visit this group at http://groups.google.com/group/structure-software.
For more options, visit https://groups.google.com/d/optout.

--
You received this message because you are subscribed to the Google Groups "structure-software" group.

To unsubscribe from this group and stop receiving emails from it, send an email to structure-softw...@googlegroups.com.

[Vinod Kumar]

I am uploading upto k=6. I am not able to interpret.

[Vikram Chhatre]

It would be useful if you put all those figures in one and create a png/jpg/pdf.

V

[Vinod Kumar]

I am putting all the figures in one pdf files....

Thnaks a lot,