The first step is for users required to input one SMILES string in the text box or draw a chemical structure on the molecular editor, that is, the input of a single molecular.
The second step is to select a fingerprint for search. Users can select fingerprints according to their demands in the drop-down menu in the first column, including both MACCS fingerprint, FP4 fingerprint, FP3 fingerprint, and FP2 fingerprint.
The third step is to set range of molecule weight. Users can set the molecular weight in the second column according to their demands. The minimum molecular weight of the compound is inputted in the left text box, while the maximum molecular weight of the compound is inputted in the right text box. The parameters in this column are not required.
The four step is to select similarity threshold. Users select the structural similarity threshold of the searched compound and the input compound in the drop-down menu in the first column according to his own demands, that is, the Tanimoto similarity between the searched compound and the input compound must be greater than the corresponding threshold, including 0.7, 0.8 and 0.9. The higher the similarity threshold, the fewer compounds that meet the requirements in theory, and vice versa.
The last step is click the "Search" button.

The first step is molecular input, including two input methods: single molecular input and molecular file input.
Single molecular input can refer to the first step of Similarity search.
Molecular file input requires users to upload a file that can input up to 1000 small molecules in TXT format. Note that this file only consists of the compounds without column titles and indexes. Otherwise, the server declares invalid input characters.
The second part is to select a molecular feature used to construct the model. Users can select CTPs, NBTPs and fingerprints to construct the prediction model, so as to predict the molecules provided by the users.

The last part is to click the "Prediction" button.