Error with Gemelli CF TypeError: '<' not supported between instances of 'str' and 'float'

Library Plugin Support
1

Hi Qiime2 folks!

I've been exploring compositionally-aware analyses and recently have been attempting to use CTF in the gemelli plugin to look at repeated measure data.

I ran the following:

(qiime2-2021.2) Mollys-MacBook-Pro-2:~ mollybletz$ qiime gemelli ctf --i-table /Users/mollybletz/My_FILES/EEID_NSF_Microbiomes/EEID_AqAdult_TempDose_Exp/16S/WithPlasmid/PlasmidWork_contamfilt_samplefilt_OTU-table_wplasmid_EEID_AqAdult_merged1234/Temp14-TS_t1-4_Filtlowsamp_CopCor_True_abund_estimate_contamfilt_samplefilt_OTU-table_wplasmid_EEID_AqAdult_TempDose_Exp_merged1234.qza --m-sample-metadata-file /Users/mollybletz/My_FILES/EEID_NSF_Microbiomes/EEID_AqAdult_TempDose_Exp/16S/Metadata_NSFEEID_16SRuns_1234Merged_plusExpData.txt --m-feature-metadata-file /Users/mollybletz/My_FILES/EEID_NSF_Microbiomes/EEID_AqAdult_TempDose_Exp/16S/WithPlasmid/EEID_16S_1234_Taxonomywplasmid.qza --p-state-column SampleDay --p-individual-id-column UTKID --output-dir EEID_14C_TimeSeries_t1-4_Gemelli_CopCorData

And I have been getting the follow error:

**Plugin error from gemelli:**

**'<' not supported between instances of 'str' and 'float'**

**See above for debug info.** 

Traceback (most recent call last):
  File "/Users/mollybletz/miniconda3/envs/qiime2-2021.2/lib/python3.6/site-packages/q2cli/commands.py", line 329, in __call__
    results = action(**arguments)
  File "<decorator-gen-558>", line 2, in ctf
  File "/Users/mollybletz/miniconda3/envs/qiime2-2021.2/lib/python3.6/site-packages/qiime2/sdk/action.py", line 245, in bound_callable
    output_types, provenance)
  File "/Users/mollybletz/miniconda3/envs/qiime2-2021.2/lib/python3.6/site-packages/qiime2/sdk/action.py", line 390, in _callable_executor_
    output_views = self._callable(**view_args)
  File "/Users/mollybletz/miniconda3/envs/qiime2-2021.2/lib/python3.6/site-packages/gemelli/ctf.py", line 40, in ctf
    feature_metadata)
  File "/Users/mollybletz/miniconda3/envs/qiime2-2021.2/lib/python3.6/site-packages/gemelli/ctf.py", line 122, in ctf_helper
    individual_id_column, state_columns)
  File "/Users/mollybletz/miniconda3/envs/qiime2-2021.2/lib/python3.6/site-packages/gemelli/preprocessing.py", line 346, in construct
    self._construct()
  File "/Users/mollybletz/miniconda3/envs/qiime2-2021.2/lib/python3.6/site-packages/gemelli/preprocessing.py", line 395, in _construct
    subject_order = sortset(mf[self.subjects])
  File "/Users/mollybletz/miniconda3/envs/qiime2-2021.2/lib/python3.6/site-packages/gemelli/preprocessing.py", line 393, in sortset
    def sortset(ids): return sorted(set(ids))
TypeError: '<' not supported between instances of 'str' and 'float'

Any thoughts on the cause of this error? I do have a slightly unbalanced design with individuals dropping out at later timepoints. Could that be causing the error

1 Like
2

Hi @Molly_Bletz.

I'm not involved in the CTF development (cc @cmartino) but it looks like there might be some mixed typing in your IDs. I've had this happen when a sample ID "looks" like a decimal (i.e. 11234.50). Maybe check your metadata file and see if this might be the case?

3 Likes
3

Hi @Molly_Bletz,

I believe @gibsramen is correct - it is likely your subject and/or sample IDs have both floats and strings. Let me know if changing the IDs did not fix the problem.

Thanks!

2 Likes
4

Thanks @cmartino and @gibsramen! And apologies for the delay in my response.

My sample IDs look like the following:
NSF14-NOVI-PM36
NSF22-NOVI-3M8
NSF22-NOVI-7M13
NSF14-NOVI-2M10
NSF14-NOVI-PM14
NSF22-NOVI-3M10
NSF22-NOVI-5M4
NSF22-NOVI-4M4
NSF22-NOVI-3M9
NSF14-NOVI-PM5
....

So I don't have decimals, but there are numbers and letters. Would removing the dashes help?

1 Like
5

Hi @Molly_Bletz,

What do the values in the subject column "UTKID" used for --p-individual-id-column look like? I think that is where the issue is occurring.

Thanks!

1 Like
6

Hi @cmartino ,

Those IDs look as follows:
UTKID
M36
M38
M40
M42
M39
M43
M44
M37
M41
M45
...

1 Like
7

Hi @Molly_Bletz,

Thank you for your patience.

Would it be possible to send the full metadata for the columns being used with CTF? I can't see why any of the IDs you sent would cause an error unless there is some exception (some missing values or mixed types).

Thanks,

Cameron

8

This topic was automatically closed 31 days after the last reply. New replies are no longer allowed.