How to filter OTU table by samples

[sampie]

Hi all!

I am using filter_otus_from_otu_table.py with --min_count_fraction to drop low abundance OTUs. However, I have combined fastq samples from two different runs which have very different sequencing depths and I would like to do this filtering so that filtering is done for each sample (in OTU table) separately. Is this possible with some QIIME tool?

Thanks

[Stefan Janssen]

Hi.

You could use filter_samples_from_otu_table.py to obtain a biom table with just one sample; then use
filter_otus_from_otu_table.py to filter for OTUs in this single sample table. At the end, combine all single-sample tables with

merge_otu_tables.py

Since the above solution comes with a lot of fiddeling and might be too tideous for a huge number of samples you could alternatively use pythons pandas module and all its cool filtering and selection capabilities. Load the biom table into a pandas DataFrame, manipulate the data as you need and write it back to a new biom table. If you are firm in python, that might be the easier way.