Spark-VS

Spark-VS enables massively parallel structure-based virtual screening (SBVS) with Apache Spark to accelerate docking-based screening of large molecular libraries against target receptors.

Key Features:

• Massively Parallel Processing: Implements distributed execution of docking-based SBVS tasks using Apache Spark to process large compound libraries in parallel.
• Scalability and Fault Tolerance: Leverages Spark's MapReduce paradigm to provide transparent scalability and fault tolerance across commodity hardware and cloud resources.
• Cloud Resource Utilization: Optimized for deployment in public cloud environments to utilize scalable compute and storage for large-scale virtual screening.

Scientific Applications:

• High-throughput Virtual Screening: Enables large-scale docking-based screening of molecular libraries to identify potential lead compounds in drug discovery research.

Methodology:

Adapts existing docking-based SBVS software to operate within a distributed computing framework using Apache Spark; benchmarked by docking a publicly available target receptor against 2.2 million compounds in a cloud environment, demonstrating a parallel efficiency of 87%.