PRISM Proteomics Research Information Storage

1
PRISM - Proteomics Research Information Storage
Management

• March 29, 2005
• Ken Auberry and Gary Kiebel

2
PRISM System Overview
3
PRISM Functions

• Collect Information
• Capture spectra files from LC-MS instruments
• Gather metadata (biomaterial, sample prep., etc.)
• Store Information
• Raw spectra files
• Analysis results files
• Lab information
• Process Information
• Analyze raw spectra with multiple tools
• Identify mass tags, peptides and proteins
• Schedule instrument runs
• Formulate reports
• Present Information
• Web interface for general user access
• Database and file access for external tools and
  systems
• Utility programs for special purpose data access

4
Proteomics Analysis Pipelines
5
PRISM Hardware

• Storage Servers
• Bulk file storage and automated file handling
• Database Servers
• Databases
• Web User Interfaces
• Data Management System (DMS)
• Mass Tag System (MTS)
• Analysis Processors
• Automated data analysis

6
Lab Information / Data Analysis

• Data Management System (DMS)
• Campaign
• Line of research
• Cell Culture
• Biomaterial source
• Experiment
• Prepared sample
• Dataset
• Single MS Instrument run
• Analysis Job
• Peptide Identification
• Peak Identification
• Total Ion Chromatogram (TIC)

7
DMS Process Automation

• Involves file handling and storage tasks and
  operating primary data analysis tools
• Performed by manager programs
• Different types for different tasks
• Multiple copies spread workload
• Coordinated through DMS tracking database
• Processing pipelines
• Architecture is flexible and scalable

8
Protein and Peptide Identification
Mass and Time Tag System (MTS)

• Uses data analysis results and metadata collected
  by DMS
• Identifies and quantifies peptides and proteins
• Selection
• Correlation
• Aggregation
• Partitions work among multiple databases and
  manager programs

9
Data Analysis Filters

• Selection criteria are used during multiple steps
• Choosing LC-MS/MS peptide hits and LC-FTICR-MS
  features for processing
• Generating quality reports
• Accumulating quality statistics
• Querying the accumulated information
• PRISM provides a central definition mechanism
• Assures consistency
• Allows results to be compared
• Filters encompass a rich variety of parameters
• Peptide identification scoring
• Calculated quality factors
• Peptide observation statistics

10
Multiple Data Analysis Pipelines

• Peptide ID Analysis Pipeline
• LC-MS/MS mass spectrometry
• Peptide database lookup
• Feature Analysis Pipeline
• Aggregation of spectral information to define
  features
• Quantitation Analysis Pipeline
• Summary of identifications and abundance
  estimates
• Report identified proteins and AMT tags
• Provide statistics and confidence indicators
• Rollup AMT tags by protein
• Average abundances of AMT tags for each protein

11
PRISM Data Model