User Tools

Site Tools


Automated pumping analysis - ideal case

Niquist sampling of terminal bulb

To check how well the pumping analysis performs, MV took data at 2x magnification. MV0032: We see 3 animals. Cropping the bottom animal in the first 210 frames, we can do a simple stack difference at 2x, and downscaled to 1x (our usual magnification.)

This shows a linescan across the terminal bulb for the 2x and the no-interpolation downsampled 1x (frame 98 MV0032). We can technically resolve the open bulb even at 1x. Even 0.5, but this is highly dependent on the focus.

Field of view considerations: For a 20 second trace (10 before entry, 10 after) we need at least 2mm of length in the field of view. We currently have at 1x

Magnification Lx (mm) Ly(mm)
1x 7 4
1.33x 5.3 3
2x 3.5 2

We can absorb 1.33x if neccesary.

Simple pixel change measure of pumping in constant velocity worms

After pharaglow tracking analysis

From the simple difference image, we can get an ok signal of pumping. It is correlated with velocity but subtracting velocity yields an ok estimate.

After pharaglow analysis, we can use a scan across the bulb to get the opening. Here, We use the mean of the whole bulb. One could do better by identifying the width at which the two lobes are most separate. This yields a very nice clean signal. The code uses the each linescan and calculates (median(line) - mean(line)). Note how clean the signal is (blue). The pixel change signal is shown in orange.

The automated analysis is dependent on peak detection. Therefore, if we increase the apparent peak height of the trace, we can get very similar results from the difference image, the anterior part of the straight image and the posterior part!

The left 3 panels show the (median(line) - mean(line) for the anterior and posterior part, the the mean difference of the image (subtracted the normalized velocity to correct drifts). All three when rescaled give very very similar pumping rates (ca. 4Hz, right upper panel.) The peak shape is also quite consistent for anterior and posterior.

Brightfield tracking

For animals in the lawn, the bulb is clearly visible at 2x magnification and pumps can be counted. However, when they move outside and faster, we loose sight of the bulb. I do not believe this is as easy as the fluorescence assay, but we can keep it in mind for single-worm imaging with a tracking stage.

wiki/documentation/monika/automated_pumping_analysis_-_ideal_case.txt · Last modified: 2020/02/21 07:47 by mscholz