weighted VS unweighted Unifrac

[sdpapet]

Hello, I think this might be an question. I read introduction about unifrac online. I would like to know if any working rules to choose weighted or unweighted Pcoa plots. Also, the continue VS discrete.

I did some analyses. My weighted pcoa plot is not same as unweighted plot.

My weighted pcoa: PC1 (60.97%) PC2 (13.14%)  some samples group together.

My unweighted : PC1(27%) PC2 (17%) no obvious groups

Should I choose weighted pcoa to report?

How about discrete VS continue. There is no difference between my discrete and continue plots.

Ben

[Luke Ursell]

Hi Ben,

The major difference between weighted and unweighted in the inclusion of OTU abundance when calculating distances between communities. I would be very suprised if you got the same graphs using both weighted and unweighted. The correct answer, therefore, is that you should use weighted if the biological question you are trying to ask takes OTU abundance of your groups into consideration. If some samples are forming groups with weighted, then likely the larger or smaller abundances of several OTUs are the primary driving force in PCoA space, but when all OTUs are considered at equal abundance, these differences are lost (unweighted).

Discrete vs. continuous only has to do with coloring of the points, and not the calculation of UniFrac distances.

As a side note, an improved PCoA visualization tool, Emperor, developed in the lab is now available: http://qiime.org/emperor/

Best,

Luke

--
 
---
You received this message because you are subscribed to the Google Groups "Qiime Forum" group.
To unsubscribe from this group and stop receiving emails from it, send an email to qiime-forum...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

[Chiang Pet]

Thank you Luke. Is there any way to control just show discrete or continue?, since it is only about color.

Ben

--
 
---
You received this message because you are subscribed to a topic in the Google Groups "Qiime Forum" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/qiime-forum/w5_vdq9OiLk/unsubscribe.
To unsubscribe from this group and all its topics, send an email to qiime-forum...@googlegroups.com.

[Jai Ram Rideout]

Hi Ben,

Just to add to Luke's explanation, discrete coloring is useful when your category of interest is categorical, e.g. Treatment A, Treatment B, Treatment C. In this case, your samples are going to form three distinct groups (either Treatment A, B, or C). 

Continuous coloring is useful if your category of interest is numerical, e.g. describes a gradient such as pH, temperature, depth, time, etc.. In this case, the continuous coloring provides a color gradient from red to blue, which matches up with the numeric range found in your category of interest (e.g. pH).

Hope this helps,

Jai

[Luke Ursell]

Both will be outputed by default, so just click on the folder of the one you want to use. I'd recommend using Emperor and having much more control over this…

Luke

[Chiang Pet]

Thanks Luke and Jal.

Is Emperor a type of software? Can you give me the website link of this software?

Ben

[Luke Ursell]

http://qiime.org/emperor/

[Yoshiki Vázquez Baeza]

Hello,

Regarding the coloring of the samples; Emperor provides the ability to switch between a discrete color mapping and a continuous color mapping "on-the-fly". Just go to the options tab and check the "Use gradient colors" checkbox.

Thanks!