search

biocore / qiime

Official QIIME 1 software repository. QIIME 2 (https://qiime2.org) has succeeded QIIME 1 as of January 2018.
GNU General Public License v2.0
286 stars 265 forks source link

file format (otu tables) inconsistency #722

Closed RNAer closed 11 years ago

RNAer commented 11 years ago

It seems sometimes the otu table files headers have to be commented out; other times they have to be uncommented. For example, the table files created from make_otu_table.py and input to PCoA plotting need headers to be commented out, but the biplot (make_3d_plots.py) needs a table file with uncommented headers. Would it be better just to require each table to have header as the first uncommented line? Firstly, this distinguish headers from comment lines; secondly, this format is more friendly to read into R tables.

gregcaporaso commented 11 years ago

@RNAer, what version of QIIME are you working with?

RNAer commented 11 years ago

the develop version from github (I haven't pulled the commits for a couple of weeks though)

gregcaporaso commented 11 years ago

They should be in biom format then - how are the files being generated?

ElDeveloper commented 11 years ago

I think this inconsistency was solved by https://github.com/wdwvt1/qiime/commit/e43b25d99e56d7df9ac4c292dcb80ee7c00afc32 https://github.com/qiime/qiime/pull/681. @RNAer can you try pulling the latest changes?

RNAer commented 11 years ago

The files are generated from other sources, that's why I found the problem; otherwise, the various qiime scripts are aware of different the comments and headers.

Thanks, Yoshiki. I'll try it

gregcaporaso commented 11 years ago

Any luck @RNAer? If you're all set I'll close this issue.

RNAer commented 11 years ago

sorry, this thread escaped my attention...

yes, the inconsistency is solved, but arguably the header line should not be commented as I mentioned above, IMHO.