General Questions for Paper
- can we transfer parameters between pathways
- how do we model the different types of regulation?
- how does the structure help vs non-structured
- compare to gene set analysis
- compare to SPIA?
- how can we relate cell lines to clinical samples?
- how can we revert disease states to normal or treatable states?
Figures List
Figure method cartoon
Figure Synthetic data and pathway comparison to SPIA (~ 4 patients)
- 3 patients with apoptosis inactivated
- One where SPIA and "naive" GM get the wrong apoptosis inference,
and requires modeling of protein expression to get the right answer
- 1 with normal apoptosis
Figure parameter transfer between pathways
Figure sensitivity analysis (parameter choice)
Figure: Show importance of structure by scrambling/randomizing pathway structures
Figure predicted activities overview heatmaps
heatmaps of pathway x sample:
- per sample GSEA
- per sample SPIA
- per sample DAGGER
Figure classifier results
- GBM
- age at diagnosis
- survival (< > year)
- Melanoma
- Gray cell lines
References/Previous Work
Irit Gat-Viks and Ron Shamir have a highly relevant series of work:
