taxa bar plot manually

Hi,
How can i do myself the taxa barplot? I don’t like the qiime2 boxplot because the repetition of colors.

Thank you

Good morning Ana,

You can make it using any program you would like!

If you export the taxa table and read counts, you could make bar charts in Excel or Google Sheets.

I make most of my graphs using R and the Phyloseq plot_bar() function.

I know some of the Qiime devs like to make graphs in Python using seaborn or Plotly.

Would you feel most comfortable making graphs using a spreadsheet program like Excel or Google Sheets, or using a language like R or Python?

Colin

2 Likes

I have done my own charts but coloring is a huge problem. Let’s say we need 40 different colors. I have generated a colormap that would use 40 different colors but, in my eyes, there are so many shades that draw too close that I can’t differentiate. Qiime2 is good at coloring, at least if the color repeats, they are not next to each other. That was a problem for me, I have solved it but the problem I mentioned about did not go away.
Also, @colinbrislawn has explained the ways you can try very well.

2 Likes

I also have problems when trying to show many, many features with many colors.

My solution is to merge features together based on shared taxonomy (Ana, I know you have asked about doing that!), then just graph the top 10 or 11 most common taxa at that level.

I don’t have a concise example of how to do this, but here is a complete example from a past paper.

Colin

2 Likes

Hi, thank you for your reply.
Excel is the easiest one but i would like to try in R.

So, when i download the csv table, the numbers are the reads ?

1 Like

Yes i am going to do that as well but i also need a bar plot with all.
I noticed that csv file from taxa bar plot hasn't the relative frequencies, and yet, in bar plot if we press each color from each column there is a percentage. So, in order to copy that i would have to do a bar plot with relative frequencies and not with that csv table?

1 Like

No, you use that CSV table either way. If you wish to display relative abundances, simply sum the counts in each sample, and then divide each feature by the sum.

2 Likes

Those numbers are counts or reads?

Thank you

1 Like

Both, right? A FeatureTable[Frequency] is a contingency table that shows how many times a feature was observed within a sample. In your case, the features are ASVs (which came from your denoised reads), so it is the number of times that denoised read was observed in that particular sample.

2 Likes

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