Extended Data Fig. 4: Behavioural state prediction and aggression validation for animals in predatory or a bacterial food context.

(a) Average state duration for animals on OP50 or on prey larvae. Statistics were performed with Mann-Whitney-U-test using Bonferroni correction for multiple tests against WT on larvae. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001; n = 93, n = 99, n = 108, n = 104, n = 79, n = 109 for pred. feed, pred. bite, pred. search, dwelling, roaming, search for WT on larvae respectively and n = 4, n = 24, n = 53, n = 67, n = 86, n = 84 for pred. feed, pred. bite, pred. search, dwelling, roaming, search for WT on OP50, respectively. (b) Transition probabilities between states ordered by hierarchical clustering for each food context. (c) Difference in transition matrices between the different contexts, scaled by the probability of each state. Note the increase in roaming-dwelling transitions, and the search behaviour on OP50, compared to a reduction of transitions into predatory behaviours like predatory search and predatory feeding.