<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />Title: C32 Using Mass Spectrometry and Multivariate PLS Wide Discriminant Analysis for Prostate Cancer Screening (Bartell)
Author: Peter Bartell
Notes: I’ve included the usual MS Word play by play narrative, a journal, Excel worksheet used to demo the Excel import wizard, the core JMP data table, as well as the original paper that was the inspiration for me developing the case.
Why Good:
- Interesting HLS case
- Full case including instructions, spreadsheet import, JMP journal and leave behind slides
- Data tables supports use of JMP Multivariate and JMP Pro Predictive Modeling capabilities
- Phase 1: Getting Data into JMP - Query Builder, Excel Import Wizard
- Phase 2: Explore and Screen - Distribution, Fit Y by X, Multivariate, Principal Components
- Phase 3: Builder and Evaluate Models - Discriminant Analysis
- Phase 4: Communicate/Share - Mass Export to PowerPoint and Word
Attachments:
- MassSpec.jrn
- MassSpecCase.doc
- MassSpecCase.pptx
- MassSpecStudy.jmp
- MassSpecStudy.xlsx
- Mol Cell Proteomics-2005-Wright-545-54.pdf
Quick update to this. Predictor Screening can be used to screen out predictors and in JMP Pro 17, you can also use Response Screening with the proteins put into the Switch role if you want to evaluate each protein separately like SNP-Trait or GWAS style analysis.
Clustering is also used to see groups of proteins that might have similar profiles for distinguishing the two groups.
Another ideal is to use Model Screening to if one is thinking about not just finding a panel of proteins (diagnostic) but also use to predict and outcome(prognostic).
