Fig. 3

The ranking performance of various versions of PubChemLite versus CompTox using the merged benchmarking set (n = 977) with comparable metadata terms. Green: without MS/MS information. Blue: with MS/MS information (includes in silico fragmentation and MoNA library scoring terms). The increase in top ranks and decrease in missing entries with newer versions shows the influence of additional annotation content in PubChem (see Section “Identifying and Filling Gaps in PubChem Annotation Content”). The script and associated data files to reproduce this plot are available on the ECI GitLab pages [41, 42]. Figure template from [43]