Briefly, the volatility P-value for a gene is computed using a statistical test that evaluates whether or not its volatility is significantly elevated or depressed compared to the rest of the genome, controlling for its length and amino acid composition. The P-value can be calculated using a randomization procedure or by direct formulae based upon a normal approximation. This web-implementation uses the normal approximation, which produces P-values that are virtually identical to those obtained by the randomization procedure.

Users who wish to obtain the source code for this calculation can do so by contacting Joshua Plotkin.