When we do have the basics, these ideas which are commonly used in proteomics could be transferred to metabolomics:
Establishing a database with common contaminants, analogue to the "CRAP" database for proteomics.
Determining the FDR for IDs with a list of true negatives (i.e. fake database entries), analogue to the reversed databases in proteomics. Check, how this problem is solved when using proteomics spectral libraries.
Advanced / automated improvement of spectral databases, analogue to algorithms like SpecraST
When we do have the basics, these ideas which are commonly used in proteomics could be transferred to metabolomics: