TADAM: Target aware drug activity model for focused library design

In this scientific poster, Selvita highlights challenges in creating target-specific libraries for disease-related targets, emphasizing the need for accessibility, relevance, diversity, and specific chemical characteristics. The challenges include the scale of large compound databases, slow progress due to a lack of activity data for the target of interest, and limitations in the diversity of known active compounds. The solution presented involves a proprietary deep learning model that predicts compound activity towards any target's pocket. This model facilitates the virtual screening of large compound libraries and provides insights into the essential parts of compounds for predicted activity.