UniFrac and the Bray Curtis distance calculation on feature table

Hi, Please let me know if I can calculate UniFrac and Bray Curtis distance by the feature table using qiime. Any commands?
Thanks!
Manoj

Hi @manojndsu,

These steps are described in detail in various online tutorials — and is a repeat of questions that you have already asked, and that a few of us have already answered. Have you checked out any of the following?

1 Like

Hi,
I am using MetONTIIME to analyze my data. I can see the distance matrix data however I am not sure why Bray Curtis distance matrix calculated from sequences is very high compared to UniFrac.

---------Bray Curtis------------

sample08 sample09 sample10
sample08 0 0.67475
sample09 0.67475 0
sample10 0.53295 0.58635

-------UniFrac----------

sample08 sample09 sample10
sample08 0 0.410938
sample09 0.410938 0
sample10 0.331465 0.351139

Hello Manoj,

I don’t think it’s too high. Bray-Curtis dissimilarities are calculated very differently from UniFrac distances so it makes sense that they are different. (If would be far more strange if they were similar!)

The range for both seems normal. I think you are good to go! :+1:

Do you have any other questions?

Colin

Hi, I have more questions: Simpsons is very high as are the others. Since Chao1 is based on singletons and doubletons in the data set, what are your thoughts on the reliability of that one with minION data?


chao1
sample08 38406.87
sample09 34957.0
sample10 36580.92

shannon
sample08 9.10
sample09 8.67
sample10 8.31

simpson

sample08 0.98512808
sample09 0.973764965
sample10 0.961517185

Good morning Manoj,

You bring up an interesting question: How do these metrics change based on sequencing platform? I have absolutely no idea, but maybe someone else does!


Hey, I’ve noticed something:

I really like that you are using multiple methods, and comparing the results. In order to see if these results are reasonable, you could compare your results to other publically available datasets on Qiita. With this as point of reference, you would know that simpson results are often around >0.9

Without the confirmation of other studies, it’s hard to know if the data is reasonable or not.
So get some new data sets and see how your data compares! :earth_americas: :earth_africa: :earth_asia:

Colin

Hi, I have no idea of using Qiita. It seems complicate of using Qiita.

At the moment I just need to know about the results of these indices. Are the results reasonable or way off.
Thanks!

:thinking:

I’m not sure… What results did you expect? What results have other researchers seen?
(You don’t have to use Qiita, but you do need a outside point of comparison.)

I gone through some research articles, however I did not see that much alpha diversity in samples. In my data Shannon shows remarkably high diversity.

1 Like

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.