(567ca) Parametric Simulations Reveal Steady-State Properties of Feedforward-Loop Motifs

We use modeling and parametric simulations to shed light to the properties of feedforward loop networks. We show that for a two-input, one output system, adding a feed-forward loop connection can improve the system's sensitivity to inputs, but only for specific activation-repression connections. A simple two-input, one-output system consists of three genes, X, Y, and Z, where input signal molecules Sx and Sy bind and activate gene products X and Y, respectively, where active product X and active product Y both control the transcription of gene Z. With the feed-forward loop connection, active product X would also control the transcription of gene Y. In our model, we have used reaction rate laws to derive the steady-state binding interactions of all proteins, signal molecules, and gene operators, and we have modeled gene expression rates as a linear function of percentage of bound operator sites. A parametric study varying inputs Sx and Sy on a logarithmic scale produce a saturating surface plot of output expression similar to experimental observations [1]. Using the two networks described (feed-forward loop and simple two-input), we present the simulation results for all combinations of activation and repression between genes, 12 total. Results show a strong correlation between relative FFL appearance in nature [2] and the input sensitivity of the motifs as compared to a simple network. Conversely, the simulation of the two rarest FFL connections shows that Sy affects the system very weakly under all conditions of Sx. We will present these results and expand the simulations to characterize each system in the presence of noise.

[1] Y. Setty, A. E. Mayo, M. G. Surette, U. Alon (2003). Proc. Natl. Acad. Sci. USA, 100, 7702-7707.

[2] S. Mangan and U. Alon (2003). Proc. Natl. Acad. Sci. USA, 100, 11980-11985.