Hi,
I have a large dataset of about 400 soil samples. I have followed the “Parkinson Mouse” tutorial; however, I cannot find any relevant patterns when I check alpha or beta diversity.
Additionally, what would be an alternative way to display the identified ASVs instead of using a taxa-barplot? When I do it for my 400 samples, the barplot is just too large.
Cheers,
Pablo
Hi!
Did you check "Moving pictures" tutorial? There are examples for alpha and beta diversity metrics
You can group your 400 samples by some parameters in your metadata file instead of plotting all 400 samples
Hi @Pablo_V,
The firstt thing to keep in mind is that tutorial datasets are designed to be easy! It's one fo the weird keys in picking/working with tutorial data. So, it doesn't surprise me that you're not seeing clear clustering. Frankly, if I see that much clustering normally in my samples, I worry about technical effects rather than real biological signaling.
There are a couple approaches I might consider. I've been using Adonis a lot recently to get a sense of relative effect sizes with metadata. This can kind of help you screen. You can also visualize with boxplots (beta-group-significance) to give you a sense of comparison.
I would suggest taking your data outside of QIIME for an area plot. R, Excel, python, stata(?), crayon (
) ... pick your poison. It's still very descriptive, but its there.
"There is no difference" is a valid answer to an experiment if you've been rigerous. (So is "not enough data").
Best,
Justine