/
IN VITRO DROPLET INSTRUCTIONS.txt
143 lines (91 loc) · 6.88 KB
/
IN VITRO DROPLET INSTRUCTIONS.txt
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
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
2019-01-20 Directions to run in vitro droplet assay analysis
Example files are in the "example" folder in the code directory (instructions below).
##### RUNNING PROGRAM #####
Open Terminal and connect to Tak4:
$ ssh -Y username@tak4.wi.mit.edu
Go to directory with python code:
$ cd /lab/solexa_young/scratch/jon_henninger/tools/in_vitro_droplet_assay
Run Python virtual environment:
$ source venv/bin/activate
It should say "(venv)" to the left of your username
To run the script, run the following function with a required argument of an Excel
metadata file. You can also add optional arguments described below.
Without options:
$ main.py path/to/metadata/file.xlsx
With options (options can be in any order):
$ main.py path/to/metadata/file.xlsx --param1 input --param2 input
When you are completely finished, either close the Terminal window, or type:
$ deactivate
Which will take you out of the Python environment.
##### REQUIRED ARGUMENTS #####
path to metadata file (e.g. /lab/solexa_young/scratch/metadata.xlsx). Path and data must be on Tak4-accessible
server (like solexa_young or young_ata4)
Metadata has the following columns (column name MUST be exact):
image_path : full path to every channel image (.TIF)
experiment_name : unique name for each sample/experiment. Items with the same
experiment_name will be grouped together as replicates
replicate : integer number corresponding to replicate
channel_id : integer number corresponding to the channel (488, 561, 642)
This will be used if the 'scaffold' parameter is set
##### OPTIONAL ARGUMENTS #####
--o : Output directory name to use. Default is the name of the metadata folder with "_output" added to the
end. In both cases, the output directory will be stored in the same directory where the metadata
is stored.
--tm : Intensity threshold multiplier. Used to identify droplets. Pixels with intensity less than
(mean_background_intensity + (standard_deviation_intensity * threshold_multiplier))
will be thresholded out. Default is 3 (so 3 std's above the mean).
--r : Subset circle area (in px^2). After droplets are detected, droplet intensity is calculated in three ways:
(1) Mean intensity of whole droplet
(2) Max intensity of whole droplet
(3) Mean intensity of smaller inner subset of droplet
The (3) option is to account for intensity drop-off at the edge of droplets. The 'r' parameter controls
the radius of the circle centered on the center of droplet to measure. Default is 9 px^2.
--min_a : Minimum area threshold for calling droplets. Default is 9 px^2.
--max_a : Maximum area threshold for calling droplets. Default is 500 px^2
--circ : Circularity threshold for calling droplets. Circularity = (4pi * Area / Perimeter^2). Generally,
this calculation falls between [0, 1], where 1 is a perfect circle (a square is about 0.6-0.7).
Because these calculations are done on pixels that don't form perfect lines,
sometimes circularity can be bigger than 1. Default is 0.8.
--s : Channel to use for scaffold. Options are 488, 561, 642, or avg. Average scaffold will take the average
of all channels. Default is avg.
--b : Value for background subtraction. Just does a straight subtraction. Default is 0.0.
--pr : Droplet measurement to use for partition ratio. Options are:
'sub' for subset circle, 'mean' for mean intensity of whole droplet, 'max' for max intensity in droplet.
Default is 'sub'
--crop : Integer number to crop the image into a smaller square (width from image center in pixels).
Useful if uneven illumination at edges is a problem. Defaults to entire image (image_width/2)
--no-image : Flag that specifies not to output images of the called droplets that are saved to output.
Will speed up the pipeline by a bit. Only need to write "--no-image"
--rand_bulk : (STILL IN DEVELOPMENT; CAN'T BE USED YET)
Flag that will calculate the bulk (C-out) differently. Intead of averaging outside the droplets,
which could include uncalled droplets or out of focus droplets, it will randomly shuffle the image
100 times and average across these random images. This should keep the same overall background but
average out the droplets.
##### OUTPUT #####
The output will contain the following directories in the same directory as the metadata file:
/individual/
/droplet_images/ : This will include pictures (.PNG) of the called droplets
to manually verify and/or tweak parameters.
individual_output.xlsx : This will contain all output data from individual droplets with unique IDs for each
droplet. Each sheet is a unique experiment/sample.
scatter plots : If there are two channels, then there will be scatter plot outputs of the droplet
intensities in each channel plotted against each other.
box plots : Box and dot plots of the partition ratios for all droplets for every experimental
condition for all channels
histogram plots : Histograms of droplet areas for all experiments. Not on same scale.
/summary/
summary_output.xlsx : This will contain all averaged output data from individual images of droplets.
error bar plots : Error bar plots of partition ratios and condensed fraction for every channel
(WARNING: Not 100% sure that condensed fraction calculations are working correctly).
output_analysis_parameters.txt : A text file that shows the parameters used for the most recent run.
This file will be written over every analysis unless you change the name.
______________TEST___________________
To make sure pipeline is working:
Open Terminal
$ cd /lab/solexa_young/scratch/jon_henninger/tools/in_vitro_droplet_assay
$ source venv/bin/activate
$ main.py ./example/metadata.xlsx --o YOUR_NAME
$ deactivate
Program should run in Terminal and give you progress updates. It should finish with exit code 0. Then,
you should see an output folder with your name in the examples folder. The deactivate takes you out of the python
environment. You don't have to deactivate if you want to keep running analyses. Only do it at the end