Table of Contents
The examples in this tutorial were done using Nanolyzer™, Northern Nanopore's data analysis software.
1. Loading your data
The first task involved in extracting fits from your data is to visualize it. Currently, in additition to supporting any aribtrary binary file format, Nanolyzer natively supports 3 preset file formats: .bin files generated by NNi acquisition software, .log files natively generated by the Chimera VC100, and .abf files. To get started, select "Load Analysis" and pick out your data files. If you are using VC100 files or NNi binary files, you can select one at random from the dataset and it will automatically concatenate and order all files with a matching base name, ordered by timestamp.
2. Visualizing your data
From there, set your filter parameters and press "Update Trace" to see your data. Be careful not to plot too large a section of time, since nanopore data is quite dense and it may take a while to render the plot. You can also calculate power spectra of the data segment of interest to get a sense of the noise parameters of your experiment and to better inform your choice of filter.
3. Filtering your data
In order to extract signals with sufficient signal-to-noise ratio in nanopore experiments it is often necessary to digitally low-pass filter the signal to remove some of the contribution of high-frequency noise. In nanopore systems the limiting noise components scale with frequency and are due to capacitance arising from the membrane and support structure and from parasitic capacitance within the sensing electronics, so cutting off the high frequencies is sometimes useful. In Nanolyzer this is done using a Bessel filter, which is specified by two parameters - cutoff frequency, and filter order. The first parameter specifies a frequency above which the signal is attenuated. The second determines how fast the attenuation happens.
While there is no hard rule of thumb for what filter will be best for your data, in general, less is more. The closer your cutoff frequency can be to the maximum value (half of the sampling frequency, also called the Nyquist frequency), the better. Ideally no filter would be used at all, which can be achieved in Nanolyzer by simply leaving the associated entries blank. The reason to avoid it is because the filter will distort your signals, turning sharp transitions in your data into smooth ones. The less you filter, the less distortion you will see. That being said, sometimes filtering is a necessity.
Data visualization is pretty simple, and there's not much to it. Once you have your data loaded and your filter chosen, we move on to actually setting up the event fitting.
Last edited: 2021-10-18