| Both sides previous revision Previous revision Next revision | Previous revision |
| usage:division_detection [2022/09/05 23:04] – pseudomoaner | usage:division_detection [2023/07/05 15:36] (current) – [The Division Detection module] pseudomoaner |
|---|
| Despite these differences, most of the machinery responsible for performing these assignments remains the same. The layout and usage of the division detection module is therefore very similar to that of the tracking module. For the purposes of conciseness, the below usage guide focuses on the ways in which the two modules differ, rather than providing a complete guide on each aspect of the division detection process. To initialise it, simply press the 'Divisions' button on the home panel. The following GUI should now appear: | Despite these differences, most of the machinery responsible for performing these assignments remains the same. The layout and usage of the division detection module is therefore very similar to that of the tracking module. For the purposes of conciseness, the below usage guide focuses on the ways in which the two modules differ, rather than providing a complete guide on each aspect of the division detection process. To initialise it, simply press the 'Divisions' button on the home panel. The following GUI should now appear: |
| |
| {{ :usage:annotatedgui.png?nolink&800 |}} | {{ :usage:divisiondetectionv2.png?nolink&800 |}} |
| |
| ===== Model training ===== | ===== Model training ===== |
| ===== Division detection ===== | ===== Division detection ===== |
| |
| Selection of the **division threshold** is also very similar to selection of the **adaptive linking threshold** in the tracking module, with the user repeatedly performing a cycle of threshold selection - division detection - validation - threshold update. However, unlike the tracking module, the relative sparseness of division events means that it is practical to perform threshold selection by repeatedly analysing the entire dataset, rather than a single pair of frames. Collating all data across all timepoints also means that a single threshold is applied for the entire dataset (rather than being varied automatically to account for changes in link ambiguity, as in the case of the tracking module). | In contrast to the Tracking module, the Division Detection module does not require the user to manually select a tracking threshold. Instead, the division detection algorithm automates selection of this threshold by iteratively modifying its value such that the resulting lineage structure is maximally connected while still remaining a tree (i.e. contains no loops). In practice, this generally gives a near-optimal balance between assigning true links - correct assignments between mother and daughter cells - and suppressing false links. |
| | |
| To perform initial division detection, simply click the **Find divisions!** button. One of two outcomes will then occur: | |
| | |
| - If division detection completed successfully, the division viewer and lineage distribution windows will update. | |
| - If division detection resulted in a lineage with a cycle (i.e. a cell marked as its own ancestor), the following warning notice will appear and the division detection process will be aborted: | |
| | |
| {{ :usage:detectionwarning.png?nolink&400 |}} | |
| | |
| In the case of scenario (2), the **division threshold** should be reduced until the warning notice ceases to appear. | |
| |
| | Once training has completed, it is therefore sufficient to simply press the **Find divisions!** button to launch this iterative algorithm and generate your lineage. |
| ==== Validation ==== | ==== Validation ==== |
| |