Table 1.

When to use which statistical test.

First Feature	Second Feature	(Non)-Parametric?	APPROPRIATE TEST
categorical	categorical	non-parametric	hypergeometric
quantitative	quantitative	parametric	Pearson
quantitative	quantitative	non-parametric	Spearman, Kendall
categorical	quantitative	parametric	two-sample t-test
categorical	quantitative	non-parametric	Wilcoxon-Mann-Whitney, Kolmogorov-Smirnov

When analyzing the statistical association between two features across the genome, the choice of statistical test depends on whether the features are categorical or quantitative, and whether or not a parametric method can be used. For each case the appropriate test is listed.