/
improvement_notes
42 lines (28 loc) · 1.95 KB
/
improvement_notes
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
from Lin, Parker et al. Molecular Cell 2015:
Background intensity subtraction
Used an image of a homogenous soultion to correct for uneven illumination.
Droplet intensities were measured by averaging the intensities at the center (with a diameter of 2.5 µm smaller than
that of the droplets) from at least three areas.
For bulk intensities, identical samples were prepared and centrifuged then imaged. It is unclear if they actually called
"droplets" in the bulk analysis.
Intensities were converted to concentrations through a standard curve. There was no info on how this standard curve
was generated.
The partition coefficient is defined as [GFP]droplet/[GFP]bulk and shown as mean +- SD from three measurements. I assume
these were 3 independent droplet assays.
from Banani, Rosen et al. Cell 2016
Background correction by substracting dark images.
Spinning disk also corrected for non-uniform illumination by normalizing to images of uniformly fluorescent samples
Identified droplets by intensity thresholds. Regions within 2 µm of the droplet edges were ignored. Also set min and
max size thresholds for the droplets.
Intensities from all included regions of droplets or bulk phases were averaged and used to calculate molecule and module
concentrations in the two phases using an intensity v concentration standard curve.
TO DO:
@Bug Need to fix the bulk random image. It's giving me strange results!
Run correlation analysis on the intensity scatter
Maybe run statistics on the droplet data automatically?
Try calculating a different [bulk] by randomizing and averaging over many images.
Add function that only re-makes graphs from already-analyzed output
Right now, I think I might have 'channels' hardcoded in the output. I may need to change this in case people don't use
488 and 561, and in that particular order.
@Bug There might be something wrong with the condensed fraction calculation. Need to check with more data and verify.
Add better titles to droplet graphs