otu_table_mc2_w_tax_no_pynast_failures.biom problem ?

[Manuela Ramalho]

Hello!

I'm using Qiime 1.9.1, and I'm having problems with the BIOM table otu_table_mc2_w_tax_no_pynast_failures.biom. When I use the command "pick_open_reference_otus.py" I can create the table. But I can not make progress with the analysis. It seems to have a problem with the table "otu_table_mc2_w_tax_no_pynast_failures.biom". As I continue the analysis with "otu_table_mc2_w_tax.biom" and everything works fine! Does anyone know help me?

Please, I'm so close !!

Thanks in advance.

All the best,
Manuela

[jonsan]

Hi Manuela, 

I responded to your post in a different thread, but if you could reply in this one it would help keep things organized! :)

Thanks, 

-jon

[Manuela Ramalho]

Hi Jon !!
First, thank you for helping me!

What happens is that after I getting these .biom table, I try to use the "summarize_taxa_through_plots.py -i OTUS97_seqprep_April2016/otu_table_mc2_w_tax_no_pynast_failures.biom -o teste2016_4", and appears the error "ZeroDivisionError: float division by zero".
It seems that something wrong with the table  "otu_table_mc2_w_tax_no_pynast_failures.biom"

But this does not occur for "otu_table_mc2_w_tax.biom".

I attached the files you asked!

Again, thanks in advance!!!

All the best,

Manu

[jonsan]

Hi Manuela,

Ok I think I have an idea of what's going wrong here. There are a few samples in the _no_pynast_failures OTU table that have zero sequences left (take a look at the 'counts/sample detail' lines in the summary file). The summarize_taxa_through_plots.py script is going to try to make proportional abundance tables (# of counts for a particular OTU / total # of counts in sample); so for these few samples, the denominator is going to be zero and you get an error. To get around this, you'll need to filter those samples out of the OTU table using filter_samples_from_otu_table.py. 

Also note that there's quite a wide range of depths across your samples, so do be cautious interpreting the results. Doing something like rarefying the OTU table to a common sequencing depth may help, as might considering low-coverage and high-coverage samples that otherwise come from similar sources, and making sure you don't have any systematic biases in sequencing depth.

Good luck,

-jon

[Manuela Ramalho]

Hi Jon,

Thanks a lot for helping me!!! I followed what you suggested, and it really worked!!!

Now came another problem! The two main bacterias colors are very similar. I researched on the Internet how to change them. But I think I do not understand right.
I was able to generate preferences (prefs) file (make_prefs_file.py). It is attached. But I do not understand what is the next step to change the colors. Can you help me?


Thanks again!!

All the best,
Manu

[jonsan]

Honestly at this point my preference is to take the summarized taxon abundance tables produced by the script and import them into a different program to make my own bar charts -- if you want to make readable publication-quality images I think even something like Apple Numbers or Microsoft Excel can be made to work much better than fiddling with the matplotlib chart from Qiime. 

However if you just want to change one or two colors with this file, you can find the relevant line and change the three-number color code. For example, here's one bacterium:

'k__Bacteria;p__Verrucomicrobia;c__Opitutae;o__Opitutales;f__Opitutaceae;g__Opitutus':('red232.119402985',(232,100,100)),

which you might want to change from red 232 / blue 100 / green 100 (a kind of greyish red color) to pure bright green (red 000, blue 000, green 255) you would change that line to:

'k__Bacteria;p__Verrucomicrobia;c__Opitutae;o__Opitutales;f__Opitutaceae;g__Opitutus':('red232.119402985',(0,0,255)),

Cheers,

-j

[Manuela Ramalho]

Thanks for the tips, Jon !!!
I guess now I'll be able to proceed with the analysis!!!

Again, I am very grateful for all the help!

All the best,
Manu

[jonsan]

Fantastic!

Good luck,

-jon