Data Intensive Computing
High Performance Data Analysis Tools for Intelligent Sensor Pipelines
Principal Investigator: Navdeep Jaitly
Challenge
Large-scale biological experiments using mass spectrometry incur a repeated acquisition and analysis of patterns in redundant compounds. As a result, a large number of replicates is required to capture in-depth information. An intelligent control system to guide the mass spectrometers could reduce the number of required experiments.
Approach
- Develop a high performance pipeline that manages 1) data as they are captured by the mass spectrometers, and 2) the processing algorithms that analyze this data on-the-fly.
- Use the MeDICI framework to tie together the different processing components control the instruments.
Impact
A significant reduction in the number of required experiments by percentage is expected. This approach can potentially be applied to other situations requiring mass input from a
Collaborations
- PNNL's Biological Sciences Division
Accomplishments
- Feature discovery algorithm implemented
- Alignment algorithm prototyped
- Prototype pipeline in development
- Plans for software development and testing on actual biological research system
Pipeline