Entering edit mode
2.7 years ago
wyj.jackson
▴
20
I want to compare two methods. Both methods try to predict cancer driver genes.
Assume that we know the ground truth that there are M genes in total and N out of M genes are driver genes. The first method predicts m1 genes and n1 out of m1 genes are driver genes. The second method predicts m2 genes and n2 out of m2 genes are driver genes. We can use the hypergeometric test to compute the p-values for methods 1 and 2.
My question is that can we say method 1 is better if the corresponding p-value is smaller?
