Using Distruct with fastStructure

[kot...@iapg.cas.cz]

Hi,

Is there a way to display fastStructure results with Distruct?

Distruct seems to require a non-empty file with population assignment probabilities (*.popq). However, fastStructure does not seem to be outputting these.

Thank you.

All the best,

Petr

[Vikram Chhatre]

Hi Petr,

fastSTRUCTURE outputs the following file:

*.meanQ

This file contains membership coefficients (Q) for all individuals in each of the K clusters.  In other words, this is your *.indivq file.  You will have to adapt it to Distruct format by making minor changes, but the core information is provided in this file.

The population membership coefficients (*.popq) can be easily put together using this data also.  Take the average of membership for all individuals in a given population in a given cluster, that is your popq membership coefficient.

Example:

pop 1

indiv#1    0.95  0.025  0.025

indiv#2    0.80  0.10   0.10

indiv#3    0.85  0.05   0.10

Convert to popq:

pop1:  0.86  0.058  0.075

Hope that is useful.

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/groups/opt_out.

[kot...@iapg.cas.cz]

Hi Vikram,

Many thanks for your useful answer to my post - it is straightforward and helpful!

I posted this question in another topic, but got no response. I was wondering if you would know if there is any known explanation why Structure seemed to perform worse (in terms of inferring the true admixture proportions) than fastStructure in the simulation study presented in the original paper describing the fastStructure method (by Raj, Stephens and Pritchard), e.g. in Figure 4.

Thank you.

All the best,

Petr

Dne čtvrtek, 13. února 2014 14:13:55 UTC+1 Vikram Chhatre napsal(a):

Hi Petr,

fastSTRUCTURE outputs the following file:

*.meanQ

This file contains membership coefficients (Q) for all individuals in each of the K clusters.  In other words, this is your *.indivq file.  You will have to adapt it to Distruct format by making minor changes, but the core information is provided in this file.

The population membership coefficients (*.popq) can be easily put together using this data also.  Take the average of membership for all individuals in a given population in a given cluster, that is your popq membership coefficient.

Example:

pop 1

indiv#1    0.95  0.025  0.025

indiv#2    0.80  0.10   0.10

indiv#3    0.85  0.05   0.10

Convert to popq:

pop1:  0.86  0.058  0.075

Hope that is useful.

V

On Thu, Feb 13, 2014 at 4:37 AM, <kot...@iapg.cas.cz> wrote:

Hi,

Is there a way to display fastStructure results with Distruct?

Distruct seems to require a non-empty file with population assignment probabilities (*.popq). However, fastStructure does not seem to be outputting these.

Thank you.

All the best,

Petr

--
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.