biom convert ValueError: could not convert string to float:

[Brett]

Hello, 

I am assessing various methods of otu clustering with a mock community of known composition that I have sequenced. However, not all otu clustering methods generate a BIOM file as output. While trying to convert my .txt otu table to a .biom table with the "biom convert" script, I get the following error: ValueError: could not convert string to float: BY.MC.6.7_10814. I tried using QIIME labels without any decimals (e.g. BY_MC_6_7_10814), but received the same error. 

The command: biom convert -i uclust_filtered_otus.txt -o uclust_filtered_otus.biom --table-type="OTU table" --to-hdf5

The full error: attached

The .txt otu table: attached

Thanks in advance.  

-Brett 

[abir...@gmail.com]

Hi, Brett,

You mentioned that you're trying to convert an OTU table to biom format, and I see that you're passing in uclust_filtered_otus.txt, but, as far as I can tell from looking at the uploaded version of this file, it is not an OTU table.  When I look at it in a text editor, I don't see the standard grid of OTU id rows and sample id columns, with count values in each field; it looks more like a list of OTUs. Can you double-check your clustering method outputs and see if there's a file in the usual "classic" OTU table format that you could use instead?

Thanks,

Amanda

[Brett]

Hi Amanda, 

I got it figured out; needed to pass the pick_otus.py result to make_otu_table.py. 

Thanks for getting back to me. 

Cheers, 

Brett