before qiime2 clustering

[guanyumachao]

Before using vsearch clustering, there was no table for --table, what should I do？

[ebolyen (Evan Bolyen)]

Hi @guanyumachao,

You will want to use the vsearch dereplicate command to create a large table of every sequence, that is what then gets clustered alongside the corresponding representative sequences. (This makes clustering also compatible with denoising if you so chose, but for the old-school workflow, you just dereplicate (and ideally do a bit of quality filtering).)

[guanyumachao]

@ebolyen

You will want to use the vsearch dereplicate command to create a large table of every sequence, that is what then gets clustered alongside the corresponding representative sequences. (This makes clustering also compatible with denoising if you so chose, but for the old-school workflow, you just dereplicate (and ideally do a bit of quality filtering

Thank you for your answer. Here's my question

1. I have multiple samples and each sample has a fasta file, how can this be compared like fastq?
2. qiime2 needs to enter --table when clustering, but I don't want to dereplicate. Can I clustering directly without --table?

[SoilRotifer (Mike Robeson)]

Hi @guanyumachao,

@guanyumachao

I have multiple samples and each sample has a fasta file

1. If all you have are FASTA files, you can separatley import all of the FASTA files into QIIME 2, then merge them together.

2. Then, take the merged sequences and follow the approach outlined by @ebolyen here.