Hi Michael,

Please allow me first to state the obvious: the degradation of organic matter by bacteria is our point of view for a bunch of processes that heterotrophic bacteria employ for feeding. It involves a high number of functional groups because of the diversity of the organic matter. 

So I would recommend first to simplify the problem. For example, focusing on transport systems.  If you are familiar with KEGG then I recommend you to use KEGG modules, which group a number of Kegg Orthologies (KO) into a single element, simplifying the analysis. Please check KEGG modules in the following ranges:

M00191 - M00239: involve KOs for the transport of several sugars and peptides

M00265 - M00283: involve KOs for the transport of sugars by PTS 

Then you probably can expand your analysis to the enzymes that "do the actual job" (oxidoreductases, transferases, etc), but the transport system is a good start in my opinion. 

I hope that helps.

Best regards.

I would recommend CAZY database, which is the gold standard for carbohydrate degradation, especially the GH families. http://www.cazy.org/

As Salvador said, the degradation of soil organic matter is something not defined by a pathway or activity... but an heterogeneous group of processes.

Either, KO, GO, EC, COG, MetaCyc or any other global classification system can be used, but you are going to have to dig on your own for a list of interest. Saying that, you may want to start setting a threshold for your gene relative abundance or similar, otherwise the metagenome can become nearly impossible to tacke.

Also, you have some other specific databases base on functions such CaZy (mention above for carbohydrates), but also MEROPS for proteases, TCDB for transporters... and there is one for peroxidases but I don't remember the name right now.

5 years later, how did you manage your question ?

Personally, in order to detect exoenzymes (that should process SOM), I think we should consider an additional "filter" : occurrence of a signal peptide. I would use this application https://services.healthtech.dtu.dk/service.php?SignalP (publication : "SignalP 6.0 predicts all five types of signal peptides using protein language models").

Few years ago, I used an older version of SignalP and, then, I compared the results with KEGG.

I think it's a superficial approach but which can offer some kind of stringency as we only adress to the enzymatics actors which perform the first SOM hydrolysis reactions.

Tell me what you think about it, I'm sure these questions are still up-to-date in 2022 and will be in the future.