How to draw taxonomy bar chart

Hello!

I am new to Hello!

I am new to bioinformatics and am trying to associate taxa data with my microbiological test results.

I tried to get hints for my questions in the forum and tutorials but I got even more complicated as I read more texts. I lost the way from the taxonomy bar chart part.

My situation: after denoising my sequences and deriving diversity indexes, and rooting, I obtained taxonomy.qza file from my data.

  1. I want to export my taxa data into excel, which contains taxa information and relative ratios in each sample.

-> How can I export my taxonomy.qza data including the information that I need? I obtained feature-table_w_tax through the use of export -> biom-add-metadata -> biom-convert functions.

  1. I want to draw a taxonomy bar chart (and diversity index graphs) using R studio

-> If there are any other codes or tutorials or instructions that I can refer to, it would be greatly appreciated.

Below are my questions yet to be solved.

  1. In the current study, only one DNA was extracted from each sample, and sequencing was performed. I wonder if alpha- and beta-diversity analysis could make sense in this case.
  • The commands such as qiime diversity alpha-diversity-significance learned in the tutorial did not work, so when I searched the forum, there were comments from experts that comparative analysis was not possible with one sample per sample.

  • So, should I do have at least triplicate sequencing data for one type of sample? (ex: like.. soil1, soil2, soil3 from site A, soil4,5,6 from site B)

  1. sequencing depth

Tutorials tell me to set the sequencing depth to take the maximum number of features in the sample while maintaining the maximum number of samples, but I wonder why this will play a big role in future analysis. The relationship with alpha-rarefaction is also confusing. In my data, the value set for sequencing depth is, for example, about 9000, but all samples are already parallel from 3000 to 4000, that is, in this case, the value of 9000 covers most of the features without any problem. Why is it important in such a case?

  1. taxonomy database

Looking at the taxonomy bar chart, there were many cases where only the class or order level was presented in the legend and the rest did not appear (i.e., some features show only out to kingdom level).
Is it because there is no matching sequence in the database?
I also wonder how to apply the up-to-date version of the database (where can I find it) for taxonomy analysis.

  1. Reliability of taxonomy assignment and the relative ratio

For example, if 20% of genus Marinobacter was detected in my seawater sample, how reliable are the type and ratio? To my knowledge, after denoising the sequences via DADA2, we get ASVs, which are almost 100% trustable sequences. Is this right?

I am sorry for my many redundant questions but any help or comment would be very greatly helpful as I have no one to ask around. bioinformatics, and am trying to associate taxa data with my microbiological test results.

Good morning @Shinhyeong_Choe,

Welcome to the forums! :qiime2:

I'm glad you are learning about bioinformatics. There is a lot to learn in the tutorials, and that can be overwhelming at first. Have you done all four tutorials (Moving Pictures, FMT, Atacama soils, and Parkinson’s Mouse)? I ask because there are lots of similarities between them and as you complete all four, you will start to see the common patterns. Some tutorials go into more detail about certain section, which will answer some of your questions.

That's discussed in more detail in the PD Mouse tutorial#alpha-rarefaction


I think the 4 main tutorials are the best resource, let me see if I can recommend some other resources for your other questions.

  • This is a question about statistical testing, and it depends on your study design (the blocking you used and the groups you want to test). Let's return to this question later!

I hope this helps! Once you have a chance to review the docs and do the tutorials, we can discuss your other questions in more detail.

1 Like

@colinbrislawn
Thank you so much for your kind and prompt reply.
Maybe I was not patient enough to study all through the tutorials and reading materials. I will follow your suggestions and ask further questions. I see common patterns among tutorials but I think I missed some details.

Sincerely appreciated!

2 Likes