The high scoring sites are defined as those that have binding probabilities (according to the PBM model, as defined in Equation 3) higher than the top 0.1 % of the binding probabilities for all possible sites for that TF (the exact percentage cutoff has little impact, data not shown).
The binding probability for a Ca2+ ion near to a mobile receptor, i.e., a mobile buffer, was modeled in the same way as for the stationary receptor, with one exception.
Figure 13 shows a representative result of binding probabilities for SRF on the Actc1 and M23768 promoters, SP1 on the Myod1 promoter, and TEAD1 on the Myh6 promoter, with and without evolutionary conservation as an additional data source.
ROC curves of the estimated binding probabilities for the proposed data fusion method when combined with (a) DNA duplex stability data, (b) DNA duplex stability data and regulatory potential, and (c) DNA duplex stability data and evolutionary conservation data.
Supplementary Figures (b) and (d) show the histograms of the predicted binding probabilities for both the old and new data fusion methods, where the parameters in Figures (b) and (d) are selected according to the whole AUC and AUC30, respectively.
From these graphs, we can see that the new method improves discrimination by assigning much smaller binding probabilities for sequences with no known binding sites (no matter whether AUC or AUC30 is used), which thus results in much smaller false positive rate.
Figure 4 shows histograms of the estimated binding probabilities for the likelihood and Bayesian methods for prior strengths M = 50 and M = 100.
Given the probabilistic nature of our method, we also found it instructive to visualize the distribution of estimated binding probabilities for positive and negative test cases.
By placing the RyRs well inside the cleft, the outermost being 30 nm from the rim, we avoided this error when binding probabilities for the receptors were calculated.
In particular,,,, are obtained by comparing the computed binding probabilities (of a sequence to have a binding site for a TF) with known binding site information from the test data set, that is, "0" (no binding site) and "1" (at least one binding site).
