Monday, April 30, 2007

Timeline to Project

Zero hour - Start:

  • Write a mock-CWT function in matlab that provides the same results as the wCoefficents as the cwt() method in R. (done)
  • Write mock getRidge() and identifyMajorPeaks() functions in matlab as a transitional test for the final functions in R
  • Make a video explaining, in layman's terms, how the MassSpecWavelet package actually works.

Week 1-2:

  • Start developing multiple re-sampling algorithms (wavestrapping, bootstrapping the raw data, and re-sampling the Fourier space) . --Will probably be done in matlab, as I'm more familiar with it than R. This will start with a literature review.
  • Perform threshold analysis on each algorithm with a small dataset (probably just the MSexample)
  • Collect MS and NMR data. (the CAMDA data set and I also may have a access to MS QC data from a nearby toxicology lab.)

Week 3-4:

  • Work out bugs from the re-sampling algorithms and write test functions (which will probably be done in R).
  • Extend the testing parameters to the larger dataset to find the particular re-sampling algorithm that works best.
  • Perform similar re-sampling analysis on NMR data. --this should be trivial with the algorithms already made sound by the MS data.

Week 5-6:

  • Port/convert re-sampling algorithm(s) to R and test again (if not done already).
  • Start development of a modified identifyMajorPeaks() method to cover overlapping peaks.
  • Start on extending the current methods to NMR (or see what Michael needs done).

Week 7-8:

  • Test the modified identifyMajorPeaks() method extensively.
  • Continue working on NMR data methods.

Week 9-Deadline:

  • Do last touch-ups on whatever is needed and convert algorithms to R if not already done. (quality focused)
  • Combine re-sampling, overlapping peaks and NMR work to Michael's project (can be done independently with each completion as well).