Normalization method and when the microorganism to be reported does not exist

I am writing this because I have a question while using the humanN3 tool.

The function table was obtained using the human3 tool.
I want to compare metacyc pathways between Control and Case using LinDA.
Can I use total count without |g__Bacteroides.s__Bacteroides_caccae in this way?

Pathway 04003186_Abundance 04008856_Abundance
1CMET2-PWY 5976.246297 4933.886741
ALL-CHORIMATE-PWY 0 60.43152681

What should I do if the path is obtained but the microorganism I want to see does not exist, or if I want to know what function the microorganism I want to see (ex s__Bacteroides_vulgatus) has?

path $SAMPLENAME_Abundance

unmapped 140.0
87.0 unintegrated
Not integrated|g__Bacteroides.s__Bacteroides_caccae 23.0
Not integrated|g__Bacteroides.s__Bacteroides_finegoldii 20.0
Not Integrated|Unclassified 12.0
PWY0-1301: Melibiose decomposition 57.5
PWY0-1301: Melibiose decomposition|g__Bacteroides.s__Bacteroides_caccae 32.5
PWY0-1301: Melibiose decomposition|g__Bacteroides.s__Bacteroides_finegoldii 4.5
PWY0-1301: Melibiose Decomposition|Unclassified 3.0
PWY-5484: glycolysis II (from fructose-6P) 54.7
PWY-5484: Glycolysis II (derived from fructose-6P)|g__Bacteroides.s__Bacteroides_caccae 16.7
PWY-5484: Glycolysis II (from fructose-6P)|g__Bacteroides.s__Bacteroides_finegoldii 8.0
PWY-5484: Glycolysis II (from fructose-6P)|Unclassified 6.0

For the first question, yes - it is totally fine to do the testing at the level of the community totals (i.e. features with genus.species stratifications). I usually do that and then, for interesting pathways, look at what species changes might’ve been driving the changes at the community level.

For the second question, if you just want to browse which pathways have been associated to a given species (not in any particular metagenome), you could look up that species on MetaCyc/BioCyc. The alternative is to make a list of the UniRefs that show up in that species’ HUMAnN pangenome, set them all to some arbitrary abundance (e.g. 1), and then run that data through HUMAnN as if it were a genefamilies.tsv input file.