Using desorption electrospray ionization (DESI) as part of an automated high-throughput system, tandem mass spectra of the compounds in a pharmaceutical library were recorded in the positive mode under standardized conditions. Quality control filtering yielded an MS/MS library of 16 662 spectra. Fragmentation of subsets of the compounds in the library chosen to contain a single instance of a particular functional group (amide, piperazine, sulfonamide) was predicted by experts, and the results were compared with the experimental data. Expert performance was good to excellent for all the cases evaluated. Substituents on the functional groups were found to exert important secondary control over the fragmentation, with the main effect observed being product ion stabilization by aromatic substitution, which was consistent across the different groups evaluated. These substituent effects are generally explicable in terms of standard physical organic chemistry considerations of product ion stability as controlling fragmentation. A somewhat unexpected feature was the incidence of homolytic cleavages, driven by the stability of substituted amine radical cations. The findings of this study are intended to lay the groundwork for machine learning approaches to performing MS/ MS spectrum → structure and structure → MS/MS spectrum operations on the same experimental data set. The effort involved and the success achieved in computer-aided interpretation, now underway, will be compared with the expert performance as described here.