I have used run_dbcan and taken the hotpep output to identify all carbohydrate active enzymes in my metagenomic sample. I have more unique EC numbers than enzymes in the output when those EC numbers are annotated with the expasy database to get the enzyme name.
I have found the problem. It relates to EC 3.5.1.80 and EC 3.5.1.25 where the Expasy database indicated EC 3.5.1.25 used to be known as EC 3.5.1.80 but BRENDA indicated they are two different enzymes. I am now a bit confused as to which database is more correct?
In more general terms I would more trust the Expasy database as those are kinda the central repo for those EC-numbers.
EDIT. So it appears there are exceptions to the one enzyme-one ECnumber rule. For multifuctional enzymes (or fused isomeres) it can happen they receive two ECnumbers, one of each of the enzymatic function in the fused protein. eg.: https://biocyc.org/gene?orgid=ECOLI&id=EG11026#tab=RXNS
Keep in mind that these receive a double ECnumber because there exists enzymes that only exhibit one of both functions (which then receive only one of the two ECnumbers) but if in some organisms these functions have been fused into a single protein, they assign both Ecnumbers.
An enzyme shouldn't have multiple EC numbers. However, there are proteins with multiple domains, each with its own enzymatic activity. In such a case a single protein can have multiple EC numbers. Also, there are likely to be cases where enzymes have a primary activity (and a primary EC number), but may also moonlight and have lower activity on other substrates.
I have found the problem. It relates to
EC 3.5.1.80
andEC 3.5.1.25
where the Expasy database indicatedEC 3.5.1.25
used to be known asEC 3.5.1.80
but BRENDA indicated they are two different enzymes. I am now a bit confused as to which database is more correct?3.5.1.80
3.5.1.25