Patient-oriented gene set analysis for cancer mutation data.
Boca SM, Kinzler K, Velculescu VE, Vogelstein B, Parmigiani G.
Genome Biol. 2010 Nov 23;11(11):R112.
http://www.ncbi.nlm.nih.gov/sites/entrez/21092299- It's basically the same idea applied to cancer instead of autism.
- However, lacking the control patients (since cancer is high mutation frequency, that would not make sense), they have to define the null hypothesis using randomly placed mutations.
- What they call "exclusivity principle" is what I usually call the "OR combination logic" (i.e. at least one perturbed gene for the pathway/gene-set to be perturbed).