Error in core_diversity_analyses.py - TypeError: No conversion path for dtype: dtype('<U32')

[Jess Mackay]

Hi, I am very new to QIIME. 

I had some samples analysed for me on MiSeq - both 16S and ITS - and BIOM tables (one for 16S and one for ITS) were made for me by a genome research company. 

I have successfully been using the 16S BIOM table in QIIME, but when I tried to run  core_diversity_analyses.py for the ITS BIOM table I received an error: "TypeError: No conversion path for dtype: dtype('<U32')." Apparently this is raised in the step: "Filter low sequence count samples from table." 

I have no idea what this error means or how to fix it. Is there a problem with my BIOM table?

The full script I used was:

core_diversity_analyses.py -i otu_table_taxa.biom -o JEM_practice_e40000/ -m mapping_file_ext_ITS.txt -e 4000 --nonphylogenetic_diversity

The input (BIOM table and mapping file) are attached, as well as the output I received (biom_table_summary, log, and new BIOM table)

Thanks

Jess

[Daniel McDonald]

Hi Jess,

"otu_table_taxa.biom" has its sample IDs encoded as unicode as well as many of the pieces of metadata. This is technically outside of the format specification for BIOM 2.1, which is the reason for the error. I've attached a version of the table with the necessary data cast to the appropriate datatype. I also verified the filter command which threw the exception now works.

Best,

Daniel

[Jess Mackay]

Problem solved! Thank you Daniel. For future reference, can you tell me how you changed the datatype?