Fig. 4. Setd5^+/− animals display autism-compatible behaviors. Pooled male-female data of behavioral tasks illustrating genotype-dependent differences.

a (i) 3-chamber social interaction test. Left: No genotype-dependent differences in social approach as measured by ratio of time spent in chamber with novel animal vs. empty chamber (1.457 ± 0.060 vs. 1.575 ± 0.175, P = 0.5274, t_1,38 = 0.6379). (ii) Center: Het mice lack normal preference for novel over familiar animal as measured by ratio of time spent in chamber with novel animal vs. familiar animal (1.573 ± 0.130 vs. 1.021 ± 0.078, **P = 0.008, t_1,38 = 3.643). (iii) Het mice lack normal preference for cagemate over novel animal as measured by ratio of time spent in chamber with cagemate animal vs. novel animal (1.599 ± 0.149 vs. 1.187 ± 0.108, *P = 0.031, t_1,38 = 2.24). b Het mice spend significantly less time in open arms of elevated plus maze (*P = 0.0112, F_1,36 = 7.146, DF = 1). c Het mice build significantly less complex nests in overnight nest-building task (genotype **P = 0.0074, Mann–Whitney ranked U = 105.5, DF = 1). d Het mice are hyperactive in open field (genotype ****P < 10⁻⁴, F_1,2 = 18.27, DF = 1). e Het mice are impaired on Barnes maze (genotype ****P < 10⁻⁴, F_1,15 = 21.8, DF = 21.8). Statistics: n = 10 total animals per sex, per genotype; t-test of pooled male-female data (a, b, d); Mann–Whitney ranked test (c); 2-way ANOVA (e). Graphs are representative of pooled male and female data for each genotype, with replicates as individual animals and error bars mean ± SEM