Searching molecular structure databases using tandem MS data: are we there yet?

Untargeted metabolomics experiments usually rely on tandem mass spectrometry (MS/MS) to identify the thousands of compounds in a complex sample. Spectral libraries used for identification are incomplete, and many metabolites remain unknown. There has been a recent development to replace spectral libraries by molecular structure databases when searching the MS/MS data of the unknown compound. Several tools have been developed for this task, including CFM-ID, MetFrag, MAGMa(+), FingerID and CSI:FingerID. These methods are already helpful for everyday metabolomics; with further advances, these methods can become indispensable tools for tomorrow's metabolomics. Here, I discuss several questions related to this task, such as: Why not wait for spectral libraries to grow sufficiently? Why evaluate methods outside their 'comfort zone'? Should we use prior information such as citation frequencies? And, ultimately: are we there yet?