export the OTU.csv file

[Yiggrite (Yiggrite Wu)]

Hi all,

I am trying to export the OTU reads as a CSV file. But I find that Most of these OTU reads are zero. Why this happened?

[timanix (Timur Yergaliyev)]

Hello!
That means that some features are found only in few samples. One can filter feature table to remove features that are found only in 2-3 samples (as example), or have total frequency less than 10 (also just example).

Or you mean that some features have 0 in all samples?

[Yiggrite (Yiggrite Wu)]

AGmice-coral-feature-table.csv (2.8 MB)
I found almost all of these features were only found in certain one sample. That's weird. Cause at least some strain should be abundant in the gut microbe sample, like Bacteriodetes.
Could you please look at the attachment, which is the exported CSV file.

Thanks

[timanix (Timur Yergaliyev)]

There is high probability that features that are found only in one sample (assuming that you have enough replicates in the same group) are contaminates. As I wrote above, I would filter features from feature table based on total frequency and prevalence.

[Yiggrite (Yiggrite Wu)]

For my control group sample, each sample has 12517 features, but around 12176 features showed 0.0, and other values are the "contaminates" as you said. There must be something wrong...I have exported the data from the moving picture tutorial but I did see so many zero in that data.

[timanix (Timur Yergaliyev)]

@Yiggrite

For my control group sample, each sample has 12517 features,

I believe they do not have such amount of features, since most counts are 0. 12517 is the total number of unique features observed across all the samples and most likely too high to represent the actual diversity.

@Yiggrite

but around 12176 features showed 0.0, and other values are the "contaminates" as you said.

Please, do not rephrase what I wrote:

@timanix

There is high probability that features that are found only in one sample (assuming that you have enough replicates in the same group) are contaminates.

Do you have enough of replicates for the same group as the sample that have a lot of unique features? What is the difference between that one sample on other samples from the same group?

And some additional questions:
Did you run all samples in one dada2 run?
Did you remove primers before dada2?
Are all the samples were prepared in the same way? Like primers, targeted region, DNA extraction.

Also, I am puzzled that in your table you have such patterns like one on the screenshot:

Is it actually the same bacterion for each repeating count? Why counts are the same?