Alpha diversity difference between qiime1 and qiime2

[Meirav Pevsner Fischer]

Hi.

I am comparing analysis of the same dataset between Qiime1 and Qiime2.

Otu table from Qiime1  after Open OTU picking and filtering OTUs below 10 reads,  was converted to a FeatureTable in qiime2 .

In addition I prepared a FeatureTable for the same data set according to Qiime2 tutorials (using dada2 denoise-paired and training the classifier).

It looks like the alpha diversity in "observed otu" or the CHAO1 matrices are greatly reduced between Qiime1 generated FeatureTable and Qiime2 generated FeatureTable. (In the FeatureTable imported from Qiime1 we get about 600 OTUs in 20000-depth reads   and in the Qiime2 FeatureTable about 150 OTUs in the same depth)

can you please help us solve this discrepancy?

Thanks

Meirav

[Colin Brislawn]

Hello Meirav,

Qiime 1 and Qiime 2 implement very different OTU picking methods by default. Given that this core step is so very different between versions, it makes sense to me that your feature counts are very different, and your alpha diversity stats are different as well.

Does that help answer your question?

Colin