Figure 4.

Schematic representation of metrics used to compute pSDC (predicted Structural Disruption Coefficients) based on RNA structure predictions for WT (black) and mutant (magenta). The data here are for the WT, and hyperferritinemia cataract syndrome associated U22G mutant of the FTL 5’ UTR. A.) SHAPE experimental data for the WT and U22G mutant UTRs revealing a significant effect of the U22G mutation on the structure of the RNA. An eSDC value of 2.3 is computed for this data. B.) sFold Minimum Free Energy (MFE) probability of base-pairing for the WT (black) and U22G (magenta) containing sequence, one corresponds to not-base-paired and zero paired. We see that the program correctly predicts changes in the 40-60 range as measured by SHAPE. C.) Probability of base-pairing computed as the sum of the rows or columns of the partition function [64]. In this case the partition function is computed using sFold Boltzmann suboptimal sampling and computing the observed frequency of base-pairing [51]. D.) Z Centroid simplification of the partition function and probability of pairing computed by summing the rows or column [51]. E.) Probability of pairing assessed as the cluster centroid structure of the most populated cluster of suboptimal structures, in this case using sFold and k-means clustering as previously described [51].

Ritz et al. BMC Genomics 2012 13(Suppl 4):S6   doi:10.1186/1471-2164-13-S4-S6