LHASA technology relies on chemistry knowledge bases written in a proprietary "chemical English" language. This language, called CHMTRN, was developed at Harvard and allows one to specify a series of chemical questions for each transform or rule in the database. These questions, or "qualifiers," are read by LHASA and DEREK at runtime, allowing the programs to evaluate the merit of each transform or rule as applied to any target molecule.

The knowledge base of transforms in LHASA has been designed to allow LHASA to look at an arbitrary input target structure and to decide, by asking intelligent structural questions, which transforms are appropriate and which are not. The rules that comprise the DEREK rulebases are written in a similar vein.

• Direct reaction validation and rating evaluation
• Easy knowledge base updates and customization
• Built-in feedback to the user
• Extensive literature references