Here is the situation. Suppose I have a patient may be at risk of some kind of disease. And this disease may be affected by 5 genes. For each gene I have a healthy-people baseline of its gene expression. Thus I can perform a z test/ t test to tell that for each gene whether the gene expression of this patient is significant. And if 2 out of the whole 5 genes are significant, then the patient may be determined at risk of this disease. And now, as the technology developed, we can detect 100 genes rather than only 5 genes and we think if 40 out of 100 genes are significant then the patient is at risk of the disease.

My question is that I have a thought that after we upscale the gene repertoire, the statistical power or effect size or sensitivity of my method would increase. But how can I show this effect? Is there any method for me to visualize or hypothesis test/criteria such as type I or type II error for me to calculate?