Parameter estimation and validation.

<p>A. Simulated data (top) reasonably reproduces experimental measurements (bottom), including noise. Simulations include two sources of noise: 1) uniform background noise with zero mean and standard deviation <i>σ</i>_<i>N</i>; and 2) noise due to temporal fluctuations in fluorescent intensity, with zero mean and standard deviation <i>σ</i>_<i>F</i>. The standard deviation of a diffraction limited spot, <i>σ</i>_<i>GFP</i> = 2 pixels, where one pixel is 80 nm. B. Kinetic scheme defining the interaction of a fluorescently-labeled myosin (GFP-S1) with an actin binding site. The scheme comes from [<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1004599#pcbi.1004599.ref010" target="_blank">10</a>], but has been modified to reflect the low myosin concentrations in the experiments [<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1004599#pcbi.1004599.ref052" target="_blank">52</a>]. C. Simulated and measured data generate similar fitting error, suggesting that we have successfully captured signal noise. The plot shows a histogram of root mean squared (RMS) error for each of 500 frames of simulated and measured data. D. By varying the calcium dependent fitting parameter <i>ε</i>, the model reasonably reproduces measurements at low and high calcium. At high calcium (pCa 4, where pCa is the negative log of the calcium concentration) Desai et al. [<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1004599#pcbi.1004599.ref052" target="_blank">52</a>], observed fewer binding events for thin filaments than for actin alone We can replicate this result with <i>ε</i> = 0.5 (top and middle). Binding was almost completely eliminated at very low calcium (pCa 8) [<a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1004599#pcbi.1004599.ref052" target="_blank">52</a>], which we can replicate with <i>ε</i> = 0.01 (bottom). Note: for A and C, scaled intensity, defined in the text, gives a single GFP-S1 an intensity of 1. For D, zero is defined as the average minimum fluorescence achieved at each pixel. Here, scaled intensity is non-zero in the absence of GFP-S1, and a single GFP-S1 increases the signal by 1 unit.