Figure 1.

Modeling mammalian transcription with linear splines. (A) Sigmoidal transcriptional response (Carey, 1998). The response is flat below a binding affinity threshold, namely, the gene activation threshold, and varies exponentially above it. It saturates at high binding energies. (B) Example of a linear spline. A linear spline is a piecewise linear function: it is zero below (above) a threshold, termed knot (ξ), and changes linearly above (below) it. E_g refers to the observed mRNA level of gene g, whereas E_gC is its mRNA level in the reference sample. Knots are related to gene activation thresholds. All genes with PWM scores S_g>ξ are predicted targets of the motif contributing to this spline, shaded in blue. (C) A schematic view of the key steps in identification of significant motifs and motif pairs using linear splines.