Skip to main content


Fig. 4 | BMC Systems Biology

Fig. 4

From: Incorporating prior knowledge improves detection of differences in bacterial growth rate

Fig. 4

Bayes factor performs more consistently than an F-test when comparing growth rates. The results of using both an F-test (as described in the methods section) and Bayes factors (with the noise level, σ, inferred) on the same data to test whether two curves with 32 points have equal or different growth rates. We test two differences in growth rate, Δ μ max. The first curve is fixed, whilst for the second curve, 50 computationally random replicates are generated for each noise level σ L . For the Bayesian analysis, growth rates are considered to be different if there is substantial evidence for this (hypothesis 3) over equal growth rates (hypothesis 2). We also perform the analysis using prior knowledge (obtained from four previously analysed curves) for both curves. Here we make use of both the Gaussian and Cauchy priors and note that in the right hand plot the two are indistinguishable. Detection when using Bayes factors may be further improved by incorporating prior knowledge

Back to article page