Achievement

This project was led by Rick Weber and Nick Lineback with the ORNL proteomics group. Identifying proteins with mass spectrometry is a key element in proteomics. However, the complexity of proteomic datasets makes protein identification more difficult as post-translational modifications are considered. The result is a frustrating bottleneck for researchers. To keep up with these demands and empower the next generation of revolutionary informatics, SCALE-IT students built Specmaster. This program uses GPUs to accelerate searches and enables parallel programming. The first implementation yielded a 2-5x speed up over current tools. Additional features that will improve performance include a device-specific tuning and heterogeneous parallel execution. With Specmaster’s enhanced performance, users will be able to achieve increased data throughput and have the capability to run searches to include post-translational modifications.