When comparing one condition with a single sample vs one condition with more samples, how the statistic (p.adj) is computed?