Align-to-tree-mafft-fasttree error

https://docs.qiime2.org/2018.8/plugins/available/feature-table/filter-features/

https://docs.qiime2.org/2018.8/plugins/available/feature-table/filter-seqs/

Hi!
I tryed to filter table with various --p-min-frequency, but got no significant decrease of table file size. Even --p-min-frequency 1000000 led to 35 mb table. How can it be?

Is it possible you have such a high sequencing depth? And what is the full command you are using?

Could you please use qiime feature-table summarize to summarize the tables before and after filtering and post both summaries here?

1 Like

Hi.
Full command: qiime feature-table filter-features --i-table table-nonchimeric.qza --p-min-frequency 100000000 --o-filtered-table filtered-table-nonchimeric.qza --verbose

table summarize before:
https://drive.google.com/open?id=16zk469BDuKx0SnKqm8mYTHvxucFt1uhM
table summarize after:
https://drive.google.com/open?id=1RAmiUFMJW_L0Ty3V9404eJhXVDizjs_R

Hi @5cr34m,
The filtered feature table is empty. All observations have been filtered out! Which makes sense with such a high minimum frequency.

I would suggest not focusing on the size of the table file (which may not reflect the actual number of observations since other information is stored in that file). Instead filter at a sane level (that is subjective, but see the paper about filtering that I liked to above for some ideas), and then filter your sequence file based on the features present in the feature table (see the --i-table input for feature-table filter-seqs).

Thanks. I thought that size matters.

I am not sure why the table size is still so large here for an empty table, but based on the summary it obviously does not reflect the contents. So I recommend using the summary for assessing filtering efficacy, rather than file size.

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.