The probability that each of the rate-length laws reproduces the experimentally observed domain size distribution given a value of .

<p>Given a rate-length law mapping chain length to folding times, if the distribution of sizes of protein domains is modeled as a Gaussian, choosing a timescale above which proteins need assistance to fold, and an approximate fraction of proteins that need assistance to fold fully specifies the distribution of domain sizes (<a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0078606#pone.0078606.e057" target="_blank">eq. 3</a>; here, ). That distribution was used as a model for the generation of the empirically observed domain size distribution reported in <a href="http://www.plosone.org/article/info:doi/10.1371/journal.pone.0078606#pone-0078606-g001" target="_blank">Fig. 1</a>,. Here, we plot the probability that the empirically observed distribution resulted from the model indicated. Note the y-axis is and divided by for clarity, so small differences on the plot are actually quite large.