beta_diversity_through_plots.py missing mapping file column

[Rachael Lappan]

Hi everyone,

I am running beta_diversity_through_plots.py to produce a PCoA plot as follows:

beta_diversity_through_plots.py -i nps_only.biom -m ../../mapping_files/mapfile_Runs1234_extra.txt -o beta_nps_raw_extra -t ../rep_set.tre --color_by_all_fields

One of the columns in my mapping file is 'depth', which contains the sequencing depth of each sample. I am trying to colour the plot in Emperor by depth just to check that there's no clustering by sequencing depth. However, depth does not appear as an option in the drop-down menu for colouring by metadata. All other columns from the mapping file are present. I thought it may be because the values are numeric, but I have had success using a mapping file generated from summarize_taxa.py that adds the taxonomic information to the file, allowing me to gradient colour by taxon abundance (also numeric).

I have attached the mapping file, biom table and phylogenetic tree I am using. I get no errors or warnings when validating the mapping file.

Can anyone help me fix this issue? I just can't work it out!

Cheers,
Rachael

[Antonio González Peña]

Emperor by default ignores all columns that have unique values, to add them just add --add_unique_columns to your command. 

[Rachael Lappan]

Hi Antonio,

Perfect, thank you! For anyone reading, beta_diversity_plots.py won't accept the option --add_unique_columns; but make_emperor.py will. I just re-ran that part of the workflow script with this option.

Cheers,

Rachael

[Antonio González Peña]

That's correct and just running make_emperor.py will be a faster solution if you already have the *pc.txt files; however, if you are going to run that pipeline for the first time you can use the QIIME parameters file to set this option: http://qiime.org/documentation/qiime_parameters_files.html

[Jennifer Lang]

Ah this was driving me crazy! The last dataset I analyzed must have had a duplicate number of reads because it came up fine but I couldn't get the column to appear this time. Now I know why.

I just changed one of the read values that was pretty close to another one to get around it. Not the best solution but it worked real quick.

I got the make_emperor.py --add_unique_columns to work.
Never got the parameter file to work. Admittedly I only spent about 10 min trying. But can you actually use the parameter file for make_emperor.py because it's not listed as a qiime script?


I wish they would change the default and allow columns with unique values because continuous values can be very informative!

Jennifer