File S1 - Inferring the Rate-Length Law of Protein Folding

Supplemental Information. Figure S1. The rate-length data from each source used, plotted together. Even though there is some overlap in reported sequences, many identical proteins have different chain lengths or folding times reported. We combined these data for our analysis, eliminating only identical measurements. Figure S2. The parameter fit for a stretched exponential, , with a free parameter. Notice how the fit is perfectly straight on a log-log plot, a characteristic trait of power laws. Figure S3. The parameter posteriors for each model are sharply peaked around their modes. Plotted here is the maximum (not the marginal value) of the posterior at various values of the key parameters or . One can see that the likelihood has a sharply peaked value along this dimension. Table S1. The KineticDB dataset with mutants included. Since there are only a few proteins with mutants, and there are many mutants for these few proteins, this database gives artificially more weight to those individual proteins. Table S2. Bayes factors comparing datasets. (PDF)