Important: Bioactivity prediction depends on the access to the CHEMBL DATABASE API which at times could be unstable due to ChEMBL database server issues. Please check for connectivity before computing.
If the API is not experiencing lag issues, computing time will fluctuate between 2-5 seconds for each valid ChEMBL ID depending on the target compound.
To predict the bioactivity value of potential drug molecules, first navigate to the 'Target Compounds' tab and select a desired target compound. Next, upload a text file containing all of the potential drug molecules. The compounds in the text file should be separated by spaces to allow for molecule identification. Invalid ChEMBL IDs will be ignored.
By utilizing machine learning algorithms, this web-application can predict the bioactivity of input drug compounds towards a specific drug target.
The web application will then generate a downloadable CSV file with the predicted values for each molecule sorted by bioactivities.
This web application was developed using the Flask framework. To facilitate dynamic presentation of submission eligibility for files, JavaScript was employed. It also manages URL redirection, file content interpretation, and animation effects. The web application's capabilities would be expanded in the future by incorporating features such as calculation of Lipinski's Rule of Five for input molecules and an expanded list of target compounds.
