Fig. 2.

STAT increases weight; cohousing transmits that effect but only in chow diet. a–d Trajectories of mouse weight through life, stratified by sex and the experimental variable (STAT vs Control, cohoused vs non-cohoused). e Weights were log-transformed and then considered in a mixed effects model that controlled for day and sex. This plot illustrates the compensation for sex and day, but the model incorporated that compensation directly as random effects, obviating the need for a z-score transformation in the response variable (log-transformed weight). f The effects determined for the impact of the STAT and cohousing on the log-transformed weight show that STAT increases weight, and cohousing transmits that effect but only in chow diet. The scale of the effects is in multiples of the average weight (e.g. 0.1 represents an average weight that is 0.1×, or 10%, heavier)