Heterogeneous Idealization of Ion Channel Recordings – Open Channel Noise

Abstract

We propose a new model-free segmentation method for idealizing ion channel recordings. This method is designed to deal with heterogeneity of measurement errors. This in particular applies to open channel noise which, in general, is particularly difficult to cope with for model-free approaches. Our methodology is able to deal with lowpass filtered data which provides a further computational challenge. To this end we propose a multiresolution testing approach, combined with local deconvolution to resolve the lowpass filter. Simulations and statistical theory confirm that the proposed idealization recovers the underlying signal very accurately at presence of heterogeneous noise, even when events are shorter than the filter length. The method is compared to existing approaches in computer experiments and on real data. We find that it is the only one which allows to identify openings of the PorB porine at two different temporal scales. An implementation is available as an R package.

Analysis of patchclamp recordings: model-free multiscale methods and software

“…Within an HMM framework, one can only incompletely determine whether the data are compatible with the underlying model assumptions. Hence, despite the above described advantages, at least in simulations and real data examples in Pein et al (2021); Bartsch et al (2019), we observed that parameter estimates based on an idealization (either obtained by model-free approaches or by the Viterbi algorithm) appear to be more accurate than direct estimates by the Baum-Welch algorithm. Gating dynamics are time-continuous processes, but for simplification, many HMM approaches underlie a time-discrete Markov chain as an approximation.…”

“…To decide whether the noise is heterogeneous, we recommend to more advanced users also the following systematic approach: if longer segments without gating events are present, one can use them to estimate the noise level. Alternatively, one can idealize the data with JSMURF or HILDE with heterogeneous noise setting and use the idealization to determine noise levels as detailed in (Pein et al 2021, Section VI-C).…”

“…In this section, we explain the statistical models underlying our methodology. For more details, see (Hotz et al 2013;Pein et al 2018Pein et al , 2021.…”

“…It misses many of the small conductance changes, since the fact that it also looks for very short events (which are not present in this dataset) slightly decreases detection power for long events compared to JSMURF European Biophysics Journal 1 3 one proceeds as for model-free idealization. Using a layered HMM on simulated filtered signals (Pein et al 2021, Sect. IV-D) as well as in real data applications (Pein et al 2021, Sect. V) (Bartsch et al 2019), we observed that the thus estimated parameters were significantly better than the parameters obtained directly by a Baum-Welch algorithm, most likely because of the applied missed event correction.…”

“…A time-discrete approximation is also implied by most model-free approaches as they allow gating events only at the sampling points. An exception is the local deconvolution approach used in Pein et al (2018Pein et al ( , 2021. Code to obtain such histograms was explained in the paragraph 'Interpreting, plotting and verification of the output' in Using our software.…”