CPSA Digest 2003

Day 3

WeOC3

Using ChemAnalytics to Support the Drug Discovery Process

Definition: ChemAnalytics™:
"Computational algorithms and information systems that strongly link chemical structure and analytical information in order to better extract knowledge, increase laboratory throughput, and disseminate results."

ACD software is used to accelerate various levels of the drug discovery process, e.g., choosing which compounds to make, validating structures and spectrum matches, detecting and identifying metabolites and impurities, and developing separations for scale-up synthesis.

Four Important Challenges

• Instrumentation heterogeneity
  • Different forms/brands of instrumentation with different data formats that may/may not be compatible with each other
• Data interpretation
  • Specific chemometric tools to accelerate process of raw data signal to conclusion about compounds tested
• Reporting
  • Spectra, chromatograms and figures are carried into third party software to add annotation, then add into an application that generates a report
• Management of results
  • Vendor neutral system

Software Tools
CODA, Compare LC-MS, and Autoassign are software tools that speed up the process of analyzing LC-MS and MS-MS data.

• COmponent Detection Algorithm (CODA) is a chemometric algorithm
  • Reported by W. Windig, J.M. Phalp, A.W. Payne, "A Noise and Background Reduction Method for Component Detection in Liquid Chromatography/Mass Spectrometry," Anal. Chem. 68, 3602-3606 (1996).
  • Provides generic chromatographic peak extraction.
  • Apply CODA to each of the datasets imported.
  • Remove noise and background.
  • Do a comparison of what¹s left, removing the peaks that are common from each.

Reporting
Reports can be created in a reproducible format and in one step. These reports are searchable by spectrum, sub-spectrum, structure similarity, Markush structure, peak(s) and whole spectrum (HQI), compound number and user data. Metabolite knowledge (structures, pathways, analytical data, assignments) can be captured in a fully searchable metabolism database. Semi-known chemical structures can be drawn, stored, and searched using advanced Markush capabilities.

Return to Proceedings »