Extended Data Fig. 4: Power Spectral Density and Allan Deviation of nanorotors.
From: The energy landscape for R-loop formation by the CRISPR–Cas Cascade complex
a, Power spectral density (PSD) of a nanorotor (black) that was torsionally constrained on the bottom side (next to the Cascade target) in absence of Cascade. The plateau seen at low frequencies demonstrated that the angular nanorotor fluctuations were considerably confined. This confinement was released when dynamic R-loops formed in presence of Cascade (see Fig. 1d, Fig. 2a). The PSD was fitted with a Lorentzian that was further corrected for aliasing and signal integration function (green, see Supplementary Discussion 1). Best-fit parameters for the shown molecule were \({k}_{\mathrm{rot}}=7.5\pm 0.1\,\text{pN nm}/\text{rad},\,{{f}}_{\mathrm{cut}}=158\pm 1\,\text{Hz}\)). b, Allan deviation (red solid line), calculated using overlapping intervals43 and root mean square (RMS) noise after filtering with a sliding average (RMS, blue solid line) of the nanorotor trajectory used in a (in rad and in bp). Dashed lines of the same colour represent theory predictions for the best-fit parameters of the PSD (see Supplementary Discussion 2). The response time of the nanorotor was τ=γ/κ=1.0 ms (purple dashed line). Horizontal dashed lines indicate the noise level at which DNA twist changes of 3 and 1 bp become resolvable assuming a SNR of 3. Intersections of this lines with measured RMS noise provide the corresponding temporal resolution to detect these twist changes of ~4.5 ms for 3 bp and ~54 ms for 1 bp.
