Why do path_abun_unstrat and path_cov_unstrat have different numbers of rows in PICRUSt2?

[Valentina Girardi]

Hi everyone,

I have a question about the pathway_pipeline.py outputs in PICRUSt2.
After running the pipeline with the --coverage option, I noticed that path_abun_unstrat.tsv and path_cov_unstrat.tsv don’t have the same number of pathways — some pathways appear in path_abun_unstrat but are missing from path_cov_unstrat.

Could someone please clarify:

1. Why do some pathways have abundances but no coverage values?
   (For example, several MetaCyc pathways appear in path_abun_unstrat but are absent in path_cov_unstrat.)

2. How does MetaCyc regroup the E.C. numbers into reactions and pathways in PICRUSt2?
   I understand it uses the default mapfiles (ec_level4_to_metacyc_rxn.tsv and metacyc_path2rxn_struc_filt_pro.txt), but I’d like to confirm how those mappings influence which pathways are included in the coverage table.

3. What exactly does MinPath do in this step?
   My understanding is that MinPath identifies a minimal set of pathways that can explain the observed ECs, and that skipping it (--skip_minpath) allows all possible pathways to be considered. Is that correct?

4. Finally, what does disabling --no_gap_fill mean in practical terms?
   I know gap-filling “boosts” the lowest reaction abundance to the second-lowest value, but I’d appreciate an example of how that affects the calculated coverage or abundance.

Thanks in advance for any clarification or pointers to documentation that explain these behaviors!