This repository contains benchmarking analysis Matlab code for minis software. It contains all the analysis code required to reproduce Figure 2 and Figures 4-15 of the minis benchmarking manuscript.
Download the code repository and add it to the Matlab path, including its subfolders. If you are using Windows, place the code repository in your system's root folder, because ABF files have long names. Subsequently, download the minis benchmarking data repository and place it inside the minis-benchmarking repository.
The analysis code in this repository also depends on the minis software source code. Therefore, you also need to download and install minis repository.
Once all the required repositories are downloaded and Matlab path is set up, it is straightforward to reproduce the figures. In order to reproduce Figure 2, simply execute the script displayEventsProtocols.m. When requested about data filtering, press OK. Figure panels will be saved inside the folder minis-benchmarking/minis-benchmarking-data/recs/p108b/traces. Panel A corresponds to the files named algorithm_performance_data__amp0p3_sAmp0p05_n4000_RT0p5_sRT2p8_rho0_thr0p1_noiseScaleFactor1_smoothWindow1p5__11_Jan_2021_13_56_48_0001_interv_12570_13350_voltage_range_-60.55_-53.45 (in PNG, EPS, and FIG file formats). Panel B is saved in files named algorithm_performance_data__amp0p3_sAmp0p05_n8000_RT0p5_sRT2p8_rho0_thr0p1_noiseScaleFactor1_smoothWindow1p5__11_Jan_2021_13_57_31_0001_interv_12570_13350_voltage_range_-60.55_-53.45.
To reproduce Figure 5, execute the script in the file named displayEvents.m. When requested about data filtering, press OK. The figure panels will be saved inside folders minis-benchmarking/minis-benchmarking-data/recs/p108b/traces and minis-benchmarking/minis-benchmarking-data/recs/p108a/traces. Files (available in PNG, EPS, and FIG file formats) corresponding to particular figure panels are listed below:
- Panel A:
algorithm_performance_data__amp0p3_sAmp0p05_n8000_RT0p5_sRT2p8_rho0_thr0p1_noiseScaleFactor1_smoothWindow1p5__11_Jan_2021_13_57_31_0001_interv_34400_34600_volt_range_-58_-53.96. - Panel B:
algorithm_performance_data__amp0p3_sAmp0p05_n8000_RT0p5_sRT2p8_rho0_thr0p1_noiseScaleFactor1_smoothWindow1p5__11_Jan_2021_13_57_31_0001_interv_41300_41800_volt_range_-58_-53.96. - Panel C:
algorithm_performance_data__amp0p3_sAmp0p05_n8000_RT0p5_sRT2p8_rho0_thr0p1_noiseScaleFactor1_smoothWindow1p5__11_Jan_2021_13_57_31_0001_interv_23600_24000_volt_range_-58_-53.96. - Panel D:
algorithm_performance_data__amp0p3_sAmp0p05_n8000_RT0p5_sRT2p8_rho0_thr0p1_noiseScaleFactor1_smoothWindow1p5__11_Jan_2021_13_57_31_0001_interv_7400_7700_volt_range_-58_-53.96. - Panel E:
algorithm_performance_data__amp0p3_sAmp0p05_n8000_RT0p5_sRT2p8_rho0_thr0p1_noiseScaleFactor1_smoothWindow1p5__11_Jan_2021_13_57_31_0001_interv_350_500_volt_range_-58_-53.96. - Panel F:
algorithm_performance_data__amp0p3_sAmp0p05_n8000_RT0p5_sRT2p8_rho0_thr0p1_noiseScaleFactor1_smoothWindow1p5__11_Jan_2021_13_57_31_0001_interv_26300_26500_volt_range_-58_-53.96. - Panel G:
algorithm_performance_data__amp0p3_sAmp0p05_n8000_RT0p5_sRT2p8_rho0_thr0p1_noiseScaleFactor1_smoothWindow1p5__11_Jan_2021_13_57_31_0001_interv_37750_37850_volt_range_-58_-53.96. - Panel H:
algorithm_performance_data__amp0p3_sAmp0p05_n2000_RT0p5_sRT2p8_rho0_thr0p1_noiseScaleFactor1_smoothWindow1p5__10_Jan_2021_23_30_07_0001_interv_38500_38700_volt_range_-62.55_-58.8.
Finally, to reproduce Figures 4 and 6-15, run Matlab script files summaryMinis.m and summaryMinis2.m. All figures produced by these analysis scripts will be saved in folders minis-benchmarking/minis-benchmarking-data/summary, minis-benchmarking/minis-benchmarking-data/summarySelect, and minis-benchmarking/minis-benchmarking-data/summary2. All the relevant figures will be picked up and displayed in the figureDraft.mlx Matlab Live Script document. The output of the Live Script document will be saved in minis-benchmarking/paper-figures folder. Note that when executing summaryMinis2.m script for the first time, make sure that line 4 set as follows
fullRun = true;
It will take a considerable amount of time to run this script the first time. However, once finished, the results are saved in decayDetectionPerformance.mat files in all three folders. Running the script the second and subsequent times, line 4 can be set as follows
fullRun = false;
If you notice that x-axis tick marks of certain figures appear to be incorrectly labeled, rerun summaryMinis2.m script with the fullRun option set to false
fullRun = false;
and adjust lines 29-35 to get the right rate of change scaling:
if strcmpi(testType{iTest}, 'test1')
adjustRates = false;
elseif strcmpi(testType{iTest}, 'testSelect')
adjustRates = true;
elseif strcmpi(testType{iTest}, 'test2')
adjustRates = false;
end
The results of statistical inference tests are saved in stats.mat file in all three folders.
Figures are produced by loading and analysing already processed data files within the minis benchmarking data [repository](insert the link here). Large part of data processing was carried out using scripts evaluateMinisPostprocessing.m, evaluateMiniAnalysisPostprocessing.m, and evaluatepClampPostprocessing.m (or evaluateAll.m). Some data processing was also carried out during data simulation using minis software. If needed, all of the data processing steps can be repeated. For more detail, see the minis benchmarking data [repository](insert the link here).