DrugTar User Guide

The DrugTar web server enables users to predict protein druggability scores using the DrugTar method. Users can input protein IDs directly or by uploading text or CSV files. The web server outputs predictions and provides downloadable documentation for further guidance. The result table displays the following columns:

• Name: Unique identifier for each protein, typically sourced from public databases such as UniProt.
• Druggability Score: The mean druggability score predicted by DrugTar, shown alongside the 95% confidence interval in parentheses. This score represents the likelihood of the protein being a drug target.
• DrugTar Prediction: Categorization of the protein as either a "Drug Target" or "Non-Drug Target." This classification is based on a threshold of 0.32, calculated using the Matthews correlation coefficient.
• State: The status of the protein according to the ProTar-II dataset. Proteins are categorized as:
  1. Drug Target: The protein is an FDA-approved drug target and was included in the training set.
  2. Non-Drug Target*: The protein is not targeted by any FDA-approved or experimental drugs but was selected as a negative sample for training and validation.
  3. Non-Drug Target: The protein is not targeted by any FDA-approved or experimental drugs and was excluded from training and validation.