Extended Data Fig. 7: Determination of the intensity thresholds for pixel classification into adsorbed and condensed states.

a, Top, example confocal image of Klf4-GFP on λ-DNA. Bottom, intensity profile and data processing steps (see Methods). b, Probability density of the logarithm of pixel intensities of background values (after background subtraction) along a line away from the DNA (cyan dashed line in a, top). Black line, fit to a normal probability density function. The background threshold (I_th-bg = 152.9, red line in a, bottom) was defined as the mean plus 3 times the standard deviation of the distribution (grey area, upper boundary). c, Probability density of the logarithm of pixel intensities, along the intensity profiles pulling together experiments in the concentration range indicated. For low protein concentrations (top), the probability density can be fitted to a normal probability density function (black line, mean μ₁ = 187.96 and standard deviation σ₁ = 237.47). For high protein concentrations (bottom), the probability density was fitted to a two components gaussian mixture model constraining the mean of the first mode to the value extracted from the low concentration distribution in the top panel (black line σ₁ = 162.50, area a₁ = 0.22 and red line, μ₂ = 2.93*10³, σ₂ = 3.52*10³, a₂ = 0.78 respectively). We define the adsorption layer upper boundary (I_th-ads = 658.5, cyan line) as the crossing point between the first and second modes, normalized to the same area independently (grey and red areas, rescaled here for representation purposes: maximum value=0.65 and 0.50 respectively). In the intermediate concentration range, an unconstrained fit to a two components gaussian mixture model rendered a low intensity component with a mean similar to the one observed at low and high concentrations (dark grey line, μ₁ = 193.63, σ₁ = 136.37, a₁ = 0.31) and a high intensity component (light grey line, μ₂ = 634.66 and σ₂ = 708.06, a₂ = 0.69)). d, The probability density pulling together all MBP-Klf4 experiments, can be fitted to a normal probability density function (black line, μ = 295.40 and σ = 151.74). b–d, The number of experiments considered in each case (N) is indicated.