Genera shared between groups

[Camila Fontoura]

Hi

I would like to get genera results shared beteen groups, rather than OTUs, and then create a venn diagram. Can anyone help me on what scripts to use to get this data?

[Colin Brislawn]

Hello Camila,

You can groups your OTUs into genera using this script:

http://qiime.org/scripts/summarize_taxa.html 

Then, to find out how many genera are shared, you can use this script:

http://qiime.org/scripts/shared_phylotypes.html 

Qiime 1 does not include a way to make Venn diagrams, but you can make them using other programs. Maybe this one?

https://www.meta-chart.com/venn#/data 

Colin

[Camila Fontoura]

Hello

Thank you for your help! I like this Venn diagram maker, I usually use venny, but it does have some limitations, for example it does not have to choose the color etc.

I've grouped the otus for pick_closed_reference.py and to know which OTUs are shared between the groups I used the group_significance.py script. Is correct?

But I would like to cite genera, not OTUs. Do you have a script that does this?

One thing I doubted is: when I make this comparison of shared OTUs between groups, many OTUs are unique. However, when I do alpha diversity statistics, there is no significant difference between the groups. Do you know any explanation?

[Colin Brislawn]

Hello again,

The group_significance.py script compares the abundances of OTUs that found in both groups, while shared_phylotypes.py counts the number of OTUs found in both groups.

But I would like to cite genera, not OTUs. Do you have a script that does this?

All these scripts count OTUs... unless you use summarize_taxa.py to turn your OTU table into a Genera table. Then all these scripts count genera! :-)  

One thing I doubted is: when I make this comparison of shared OTUs between groups, many OTUs are unique. However, when I do alpha diversity statistics, there is no significant difference between the groups.Do you know any explanation?

Yes! What you are seeing is that the alpha diversity between groups is the same, but the beta diversity is different. 

Read about this differance here: http://www.metagenomics.wiki/pdf/definition/alpha-beta-diversity  

Great questions Camila!

Colin

[Camila Fontoura]

Thank you for the explanation. Now I understand the differences.

But for example: in the shannon index, the higher the index the more uniformity of species? or are there more dominant species than others?

And the other indexes. Chao1, simpsom, observed_otus? Do they all show the same thing?

I ran shared_phylotypes.py but it generates results per sample. I would like to get this result by group. How do I do? Can you help me? And in the exit file it does not generate the taxonomic classification, should it not contain this information?

[Colin Brislawn]

Hello Camila,

All the alpha diversity metrics measure slightly different things, in slightly different ways. Read more about that here: 

http://www.evolution.unibas.ch/walser/bacteria_community_analysis/2015-02-10_MBM_tutorial_combined.pdf

I would like to get this result by group.

OK. Then you first have to merge your samples based on group.

http://qiime.org/scripts/collapse_samples.html 

Colin

[Camila Fontoura]

Hi Colin,

Thanks!

I merged the samples based on the groups and then make the shared_phylotypes.py. But I'm having trouble filling the diagram. In the diagram, there is a field

Set 2 ^ Set 3 ^ Set 4 but I do not know how to put this information. My output file was:

Column1	SDE	FDS	FDE	SDS
SDE	1459	1119	891	1163
FDS	1119	1392	892	1104
FDE	891	892	1051	867
SDS	1163	1104	867	1448

Can you help me?

[Colin Brislawn]

Good afternoon,

I'm glad you got shared_phylotypes.py working! That results looks promising. 

I'm not sure how best to calculate these shared groups you described. It looks like they are not included in the shared_phylotypes.py script. :-(

Maybe you could color-code this table and present it as a heatmap? I know it's not the Venn diagram you were hoping for, but I think it's still a valid way to present this result. 

Colin

[hg.maryo]

Collin,

Thanks for the thread.

Is this the valid valid way you meant to represent the Camila's shared genera numbers data instead a Venn diagram?

What do you think?

Mario Hg.

[Colin Brislawn]

Hello,

Yes, that's what I had in mind. This looks good! 

(My only advice is to make sure the labels are in the same order on both sides that way the heatmap is symmetrical.) 

Colin

[Camila Fontoura]

Hi Collin and Mario,

Thanks for your help!

How did you make this chart? How would I generate symmetrically?

Another thing....I would like to generate a heatmap with several variables. As the attached picture. Can you help me please?

[Colin Brislawn]

Hello again,

I think that plot was made outside of Qiime 1, maybe using a program like R. 

Here is a tutorial on how to make heatmaps using R and phyloseq:

https://joey711.github.io/phyloseq/plot_heatmap-examples.html 

Colin

[hg.maryo]

Thanks Collin,

I was able to make it better now.

Dear Camila,

To plot the later figure I used the pheatmap package in R.

It has a lot of documentation to follow and you can custom your heatmap almost as you wish.

I found this tutorial very useful: https://davetang.org/muse/2018/05/15/making-a-heatmap-in-r-with-the-pheatmap-package/

Saludos.