Hypergeometric analysis.

This method was designed for identification of down- and up-Regulated objects
(clones and genes).
Let us denote:
Y = {Y₁, ...,Y_m} — are the measurements of the gene expression for m patients.
Where:

is ratio of actual and control signals.
Will check hypothesis H₀: Y_i = BV ∀i
BV (Boundary Value) — is designed for discrimination between up- and
down-regulation.
As a critical statistic we use:

In case of H₀ hypothesis is true, our statistic will be distributed like hypergeometric
random value with parameters N, M, n, m, where:
n — total number of clones(genes) which expression is measured in expeiment.
M — amount of values which are greater or equal then Y_i
N = mn — all expression measurements in the table.
m — total number of patients.
We suppose that for almost all genes in the table H₀ hypothesis is true.
So, for every statistic we estimate temporary P-value as:

As a final P-value for clone (gene) we take: