Which table (after denoiseing or clustering) is better to be used to create otu table?

[kaile]

There are two ways to create otu table-using the table after "denoise-singe" or "vsearch cluster-features-de-novo" function. I am confusing which way is better to be applied to further analysis. Does anyone have good suggestions?

[ebolyen (Evan Bolyen)]

Hi @kaile!

There are good reasons to cluster your data after denoising and good reasons to leave it alone. I think you would be interested in this recent topic, which covers a lot of ground related to your question:

To cluster or not to cluster?

Hello - I would like to bring up a topic that has been discussed before (https://forum.qiime2.org/t/a-quick-clarification-on-clustering/1237?u=shira) and provide an example from my own analysis. I am looking at one particular taxonomic group within environmental samples of bacterial populations. The method of denoising is dada2, so I am assuming that OTU clustering was not necessary since sequencing errors are accounted for during the denoising process. However, this analysis resulted in many O…

[kaile]

Thanks so much for your clear clarification!