Regarding Network Analysis

I have prokaryotic data of three different variants, within seven different samples of the 7 possible combinations of them (each variant alone, by pairs, or the three at the same time). I have the KEGG Orthology for each of the variants and Metabolic HMDB ID (HMP ID) and KEGG ID data for each sample. These Metabolite Data share some of the metabolites but most of then are different (as they come from METLIN data, not all of them have the KEGG IDs or the HMDB IDs)
Is it normal that when comparing the KOs with the Metabolites via Network Analysis the results of the Gene-Metabolite Networks are almost the same, even if their metabolites are different? The only metabolite compound Gene-Metabolite Network Analysis seems to recognize is the Adenosine.
I adjoint sample KOs, HMDB IDs and KEGG IDs.
A1 vs A12_HMDB_Metabolites.txt (370 Bytes)
A1 vs A12_KEGG_Metabolites.txt (167 Bytes)
LA1_KOs.txt (10.0 KB)