The Pharmit Backend: A Computer Systems Approach to Enabling Interactive Online Drug Discovery

Area of Science:

• Computational chemistry
• Cheminformatics
• Drug discovery

Background:

• Structure-based drug discovery workflows require efficient searching of large compound libraries.
• Existing methods may face limitations in speed and scalability when handling millions of molecules.

Purpose of the Study:

• To describe the system-level implementation of Pharmit, an open-source online resource for drug discovery.
• To detail the novel algorithms enabling rapid screening of large compound libraries.

Main Methods:

• Pharmit utilizes a system design focused on maximizing parallelism and minimizing communication.
• Optimized data layout for sequential processing and efficient memory management are key.
• Sub-linear time search algorithms are employed for rapid molecule screening.

Main Results:

• Pharmit enables interactive searching of millions of compounds.
• The system successfully screens millions of molecules in seconds.
• The implementation demonstrates efficient handling of cheminformatic data.

Conclusions:

• Pharmit's system design and algorithms provide a scalable solution for structure-based drug discovery.
• The resource facilitates rapid identification of potential drug candidates.
• Future directions may involve further optimization and expansion of capabilities.