Fig. 4

Model selection using BayFish and information criteria. We applied the Bayesian information criterion (BIC), the Akaike information criterion (AIC), and the deviance information criterion (DIC) metrics to the BayFish results obtained with the different parameter-stimulus models listed on the x-axis. All models were run on the same synthetic smFISH data. The maximum likelihood observed in each BayFish run was used for BIC and AIC metrics, and the full likelihood and Bayesian posterior distribution, excluding the burn-in period, were used for DIC. Models with the lowest BIC and AIC scores (left, y-axis) and DIC (right, y-axis) are the most informative models with the fewest parameters. a BayFish results for synthetic smFISH data (n=100 cells per time point) generated for an underlying k ₁-stimulus model. b BayFish results for synthetic smFISH data (n=100 cells per time point) generated for an underlying (k ₁,k ₀,μ ₁)-stimulus model. AIC Akaike information criterion, BIC Bayesian information criterion, DIC Deviance information criterion