Design Types

Design types

The Design type setting on the review step determines how group labels are interpreted.

Setting	Use for	Effect
Auto-detect	Most cases	Numeric if every group label is a number; categorical otherwise
Time course / dose–response (ordered numeric)	Time points, doses, concentrations	Numeric x-axis spaced by value; group means connected by lines; Spearman trend statistics
Categorical (genotypes / treatments / conditions)	Genotypes, treatments, cell types	Categorical axis; no connecting lines; no trend statistics

Why the distinction matters

A trend statistic assumes that groups have a meaningful order and spacing. Knockout lines or treatments have neither, so fitting a trend across them produces a number without meaning. Conversely, treating time points as unordered categories discards information.

If numeric labels are codes rather than quantities (for example genotype IDs 1, 2, 3), choose Categorical explicitly. The review step previews the consequences of the current choice before analysis.

Group order

For categorical designs, groups can be reordered with the arrows in the sample bar. The first group is the default baseline for fold changes and pairwise comparisons.