NPC Analysis

ROIManager › Workflow · version 1

Finds the nuclear pores of the file, counts their corners and localizations, and fits the labelling efficiency, in one run.

What it does

The labelling efficiency of nuclear pores takes three steps in the ROI manager: find the pores, count each one's corners and localizations, and fit the efficiency to all of them. This runs the three in one go, on the file the ROI manager is showing, and shows the results of all three in one window.

It is a chain – the plugins it runs, one after the other, each with its own settings in a section of this panel – rather than a plugin of its own: NPC, NPC Corners and NPC Labeling Efficiency. Their pages explain the methods; this one only what running them together changes. Because it is a chain, it can be edited (edit steps), saved under another name, and run over many files in a batch.

Use it for the standard analysis, and the three plugins separately to look at one step more closely – the ROIs this one makes are ordinary ROIs, and can be walked through, judged and evaluated again in the ROI manager. It needs the localization precision (xy_err_nm).

How it works

1. Group. The layers step switches the first layer to grouped, one localization per blink, as the model expects; its filter stays as the Render tab has it.

2. Find the pores. The file's earlier NPC ROIs are removed (replace earlier pores is ticked in the find the pores section), and the pores are found with the settings of that section. ROIs drawn by hand, or made by another segmenter, are left alone.

3. Count. NPC Corners is run on every ROI the manager includes, with the settings of the NPC Corners section, before the next step starts: each pore's corners seen with the precise localizations, and its localizations. The numbers are stored with the ROIs, as an evaluation in the ROI manager stores them.

4. Fit. The labelling efficiency and the blinks per copy are fitted to all the counted ROIs with the settings of the labelling efficiency section.

The window shows the pores found, the segmenter's checks and the fit (the corner and localization histograms with the model).

In detail

An evaluator in a chain. A plugin that measures one ROI at a time (NPC Corners here) runs, as a chain step, over every ROI the manager includes – every one that is ticked use, whoever made it – and the chain goes on only when all are done. A ROI on which it fails (no localizations, say) loses its own row and is counted in the step's line, not the run.

The evaluation window is left alone. The counts are stored with the ROIs as an evaluation called NPC Corners, like any other, and the evaluation window's pipeline stays as it was set up – for this work or for another. A result belongs to its ROI, whoever ran it, so the site table shows the chain's counts beside the pipeline's numbers, and NPC Labeling Efficiency run again from the ROI tab finds them, with the cutoffs they were counted with. Running the chain again replaces its counts; to keep a second set beside the first – other cutoffs, to compare – rename the NPC Corners step (edit steps). The chain's analysis reads the chain's own counts, whatever else is on the ROIs; run from the ROI tab afterwards, it asks which evaluation to read when there are several.

Nothing changes until the end. The chain runs on a copy of the session, ROI manager included; the ROIs it found, the evaluation run and the pipeline reach the ROI manager only when the whole chain has succeeded. A step that fails leaves the ROI manager as it was.

Re-running. With replace earlier pores off, the pores of an earlier run stay, a new run adds only what is new (a candidate near an existing ROI is suppressed), and all of them are counted. Results that are still current – the same ROI, data and settings – are carried forward rather than measured again.

Output

• The text: one line per step – the grouping, the pores found among the candidates, the ROIs counted, then the labelling efficiency and the blinks per copy with their errors, as NPC Labeling Efficiency gives them; on a simulation also the simulated efficiency and the truth.
• The figures: the pores found (fitted circles, kept in blue, rejected in grey), the segmenter's checks, and the fit.
• The ROIs of the pores, with their counts, in the ROI manager.
• The log: one entry for the chain, with every step's settings.

Differences from SMAP

SMAP has no single counterpart: its NPC analysis is run as segmentNPC, then an evaluation with NPCLabelingQuantify_s in the ROI manager, then NPCLabelingEfficiency (Ries 2020).

References

• Ries J. SMAP: a modular super-resolution microscopy analysis platform for SMLM data. Nat Methods 17, 870 (2020). doi:10.1038/s41592-020-0938-1

Parameters