Predict targets potentially affected by a given chemical structure.
Enter a chemical name and press "Search" to obtain search results and predictions. The system will search the
PubChem database for (i) experimental data of the query compound and (ii) structurally similar compounds (neighbors) and their experimental data. Predictions for the query compound are derived from the neighbors experimental data using an automated
read-across algorithm.
The results page is organized according to the following dimensions:
-
Gene/protein targets vs. other assays:
-
PubChem Assays with known Gene/Protein targets are listed in the first column, assays without known
targets are in the "Other assays" column.
-
Active vs. inactive targets/assays:
-
Targets/assays with positive outcomes are listed as targets/active assays, negative outcomes are
filed under non-targets/ inactive assays.
-
Experimental vs. read across:
-
Experimental data for the query compound is labelled with "experimental", read-across predictions
are labelled with "read across". In case of insufficient data no results will be displayed.i
-
Similar compounds
-
Detailed information about the compounds used for read across predictions (and their experimental data)
Source code for the project can be obtained from Github.