gavinmdouglas
commented
1 year ago Hi there,
Yes, it's possible that those pathways were filtered out, either because they were dependent on too few reactions, or due to being added to a later version of the MetaCyc database then that file was based on.
If you know all the reactions involved in the pathways you are interested in (e.g., based on the MetaCyc website), then you could add them as additional lines to that file (although capturing all the redundancy, key vs optional reaction information would be confusing). I have to admit that I took this file from HUMAnN2, so I'm not sure exactly what workflow they used to generate it originally. However, at the very least, you could add in all reactions involved, simply with equal weighting (i.e., separated by spaces).
Cheers,
Gavin
kaw97
I inherited amplicon sequencing data on wild vs managed cranberry bogs and am using picrust2’s predictions to make hypotheses for a future shotgun sequencing experiment.
I was surprised to not see any predicted pathways directly related to nitrogen cycling in my dataset (although there were others connected to nitrogen cycling, like the urea cycle and certain amino acid pathways). The nitrogen cycle is one of the things we are most interested in because of the heavy ammonium fertilization of cultivated cranberry bogs.
On further digging,
pred_metagenome_unstrat.descrip.tsvincludes EC 1.7.1.15 (nitrite reductase NADH), which maps to RXN-13854 inec_level4_to_metacyc_rxn.tsv. However, when I search for 13854 inmetacyc_path2rxn_struc_filt_pro.txtit does not appear in the file.Similarly, EC 1.7.99.4 was predicted in my dataset (nitrate reductase), which maps to Rxn-6369, Rxn-16471, and Nitratreduct-Rxn. None of these appear in
metacyc_path2rxn_struc_filt_pro.txt.Additionally, when I look at
metacyc_pathways_info.txtI see many (>10) related to nitrogen assimilation/dissimilation. However, many of them, like PWY-5674 (nitrate reduction IV, dissimilatory), do not appear inmetacyc_path2rxn_struc_filt_pro.txt.Am I correct in interpreting this as certain pathways not being covered by the picrust pipeline, or possibly certain enzymes not being linked to all metacyc versions of the reaction? EC 1.7.1.15 on metacyc includes Neurospora crassa NC-nit-6, which is also annotated as 1.7.1.4. 1.7.1.4 is included in PWY-5675. If so, is there a straightforward way of systematically adding the missing connections?