-
Notifications
You must be signed in to change notification settings - Fork 0
/
sample_data.html
executable file
·129 lines (81 loc) · 12.4 KB
/
sample_data.html
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
<!DOCTYPE html>
<html>
<head>
<meta name="description" content="HRMAn - Host Response to Microbe Analysis" />
<title>HRMAn - Host Response to Microbe Analysis</title>
<link rel="shortcut icon" href="favicon.ico">
<!--<script src="ga.js"></script>-->
<!-- Global site tag (gtag.js) - Google Analytics -->
<script async src="https://www.googletagmanager.com/gtag/js?id=UA-111014695-1"></script>
<script>
window.dataLayer = window.dataLayer || [];
function gtag(){dataLayer.push(arguments);}
gtag('js', new Date());
gtag('config', 'UA-111014695-1');
</script>
</head>
<body>
<!--<textarea theme="spacelab">-->
<xmp theme="cerulean" style="display:none;">
###[__Home__](index.html) : [__News__](news.html) : [__HRMAn2.0__](hrman2.html) : [__Sample data__](sample_data.html) : [__Downloads__](downloads.html) : [__Tutorials__](tutorials.html) : [__About us__](About_us.html)
###Example analysis of cells infected with *Toxoplasma gondii*
*Toxoplasma* is an apicomplexan parasite that can infect any nucleated cell and will establish its own vacuoles within a cells cytosol, also known as the parasitophorous vacuole (PV). Within this vacuole *Toxoplasma* will replicate until eventually the cell bursts and releases the highly infectious and replicative version of the parasite called Tachyzoite. Cells usually fight infection with intracellular pathogens by a plethora of different mechanisms. One of them is decoration of PVs with ubiquitin and subsequent recruitment of the autophagy adaptor protein p62 (Clough et al., 2016). Depending on the cell type this happens more or less frequent and has different outcomes: In HUVEC cells ubiquitylation will result in acidification of the vacuoles which indicates shuttling of the PV into the endo-lysosomal pathway, whereas in HeLa cells ubiquitylation results in autophagy-based clearance of infection (Selleck et al., 2015). Furthermore, these mechanisms seem to depend on the strain of *Toxoplasma*.
To quantify the dynamics of infection and the levels of ubiquitylation of PVs HRMAn uses sophisticated segmentation algorithms and neural network-based image analysis. The following text will illustrate how all readouts are created and how they are useful for the researcher to answer important questions in the field of host-pathogen interaction.
To illustrate how HRMAn can be used to quantify infection of tissue culture cells using high-content imaging, we infected HeLa cells with the type I *Toxoplasma gondii* strain RH. The cells were seeded on glass-bottom 96-well plates and treated with 100 IU/mL IFNγ or left unstimulated overnight. The next morning, we infected them with an MOI = 1 of RH *Toxoplasma*. 6 hours post infection the cells were fixed with paraformaldehyde and then stained for p62 and Ubiquitin using a standard immunofluorescence protocol. The parasites used for infection were genetically modified to stably express GFP.
Images were acquired with 40x magnification on a ArrayScan™ XTI Live High Content Imaging Platform. Once the pictures were acquired they were loaded into the HRMAn analysis pipeline and analysed.
The first thing HRMAn does, is to organize and pre-process the data. This includes an illumination correction and clustering of the data into sample groups as defined by the user. This is achieved by loading a plate map into the program that tells the analysis pipeline which wells represent replicate conditions. This information will be used to calculate all readouts and do the statistics for the user. Once this is done the images are automatically segmented into labels of nuclei, the cells or the pathogens and each label is assigned a unique ID. With these labels, as simple geometrical objects, the computer can work easier and calculate their sizes, shape descriptors and also the dependencies between different labels, e.g. how many parasite labels are contained within a cell label.
The easiest readouts created from an experiment like this is __Percent infected cells (%Infected)__ and __Infection levels__. Percent infected cells will show which proportion of cells contain a pathogen and infection levels will reveal the distribution of infection. As one cell can contain more than one pathogen the infection levels are a more detailed measurement and as you can see in the figure below, while IFNγ does not have an influence on the proportion of infected cells, it leads to killing of pathogens and thus more cells contain a lower number of pathogens.
As a readout for pathogen replication HRMAn can simply determine the average size of pathogen containing vacuoles __(Mean vacuole size)__. This is a bulk measurement for replication of the Toxoplasma parasites. To predict how many parasites are contained in each individual vacuole, HRMAn can read an annotated dataset and use this to train a decision tree classifier. Using this, the unknown vacuoles can be classified based on how many parasites they contain __(Replication)__. To simplify this information HRMAn also calculates the proportion of vacuoles that contain replicating parasites __(%Replicating)__.
Now that we know the infection levels and the parasites that replicate we can create even more detailed readouts. One is the __Vacuole to Cell ratio__. This normalises the number of imaged vacuoles to the number of cells and gives a quite simple measurement for killing of pathogens. This could already be observed when looking at the infection levels. Combining the number of vacuoles with the replication data we can also calculate the overall number of parasites and plot the parasite to cell ratio, or __Parasite load__. This is a combined measurement for killing and replication restriction of pathogens.
Taken together the results above show, that HeLa cells kill and restrict the growth of type I RH *Toxoplasma gondii* in an IFNγ dependent manner.
Another measurement HRMAn provides for the user is the vacuole position. This is using the centroids of the vacuoles and of the nucleus of the cell containing the PVs. Using a simple Pythagoras, the __Euclidian distance (Vacuole position)__ can be calculated. In our example experiment, IFNγ does not have an influence on the position of the vacuole inside of a cell.
Overall these are the readouts HRMAn will produce for a simple infection analysis, but as we stained the cells for proteins that are being recruited to PVs we can use HRMAn to quantify this as well.
Shown in the picture below are some example vacuoles that display recruitment of protein to them or not.
Using a neural network that can be trained for any staining and pathogen, HRMAn is able to crop out the vacuoles and then classify them based on whether or not they are decorated with Ubiquitin or p62.
With this information we can now assess how many vacuoles are decorated and we can see that HeLa cells do not decorate many vacuoles without IFNγ treatment and also do not decorate many (max. 30% in total) with IFNγ treatment __(Recruitment)__. Most of them show co-recruitment of both, Ubiquitin and p62 to the vacuoles, which is in line with previous findings based on manual counting. The biggest difference is, that we analysed more than 25,000 vacuoles in one experiment, instead of just counting 100 vacuoles per condition as it was usually the case in the recent literature.
Also, with this information we can further analyse the response of individual cells. As seen in many publications not every cell responds equally to infection and this heterogeneity can be quantified based on whether a cell decorates vacuoles or not __(%Cells decorating)__. This reveals that IFNγ increases the proportion of cells that show this behaviour. Also, if a cell responds by decorating vacuoles, they usually do not do this to all the PVs that they contain. The proportion of vacuoles that can on average be decorated per cell is another readout to quantify an individual cell's response __(%Decorated vacuoles per cell)__.
Now, that our data is partitioned into two parts - recruited and non-recruited vacuoles - we can apply all the other measurements for parasite replication as done before for all vacuoles and investigate the effect of decorating vacuoles with Ubiquitin and p62.
This reveals that parasites contained in vacuoles that are decorated with Ubiquitin and p62 are on average less replicative as all vacuoles and this can be quantifies based on the proportion of replicating parasites __(%Replicating recruited)__ and also using the decision-tree classification again give the individual number for each group __(Replication recruited)__.
Going back to more classical image analysis ways, we can then calculate the radial fluorescence intensity surrounding each decorated vacuole and extract more information of the coating itself.
First, we can define the maximum fluorescence of the coat and measure its distance from the centroid of our pathogen. This gives us a measurement for the average distance of the coat __(Coat distance)__. Also, we can measure the average fluorescence intensity of the coating and use this as a readout for the coating thickness or more precisely for a quantification of the amount of protein located at the PV __(Coat intensity)__.
Take together these results show that treatment of HeLa cells with IFNγ increases the proportion of cells that do respond to infection by decorating vacuoles, the number of vacuoles that each cell decorates increases and that the amount of protein at each decorated vacuole increases. Furthermore the analysis shows that mots vacuoles are co-decorated with Ubiquitin and p62 and that p62 is on average further away from the *Toxoplasma* centroid than Ubiquitin, which gives further evidence that Ubiquitin is first recruited to the vacuole and then p62 binds to it with its Ubiquitin binding domain and then induces further host response mechanisms that include growth restriction of the parasites and killing of them which together fight infection.
So, from one single experiment HRMAn can give you the following results:
- %Infected
- Infection levels
- Mean Vacuole size
- Replication
- %Replicating
- Vacuole to cell ratio
- Parasite Load
- Vacuole position
- Recruitment
- %Cells recruiting
- %Vacuoles recruited per cell
- Replication recruited vs. non recruited
- %Replicating recruited vs. non recruited
- Coat distance
- Coat intensity
All of which the program will automatically determine the means and calculate the statistics based on the user-defined sample groups. If you now think that you can fit 20 different samples onto one 96-well plate (conditions done in triplicates), HRMAn is a very powerful tool to study host-pathogen interaction. If you have any further questions on how it exactely works and how you can use it for your research, please feel free to contact us.
__References:__
[Clough, B., Wright, J. D., Pereira, P. M., Hirst, E. M., Johnston, A. C., Henriques, R. and Frickel, E.-M. (2016). "K63-Linked Ubiquitination Targets Toxoplasma gondii for Endo-lysosomal Destruction in IFNγ-Stimulated Human Cells." *PLOS Pathogens*, 12(11), e1006027.](https://doi.org/10.1371/journal.ppat.1006027)
[Selleck, E. M., Orchard, R. C., Lassen, K. G., Beatty, W. L., Xavier, R. J., Levine, B., Virgin, H. W. Sibley, L. D. (2015). "A Noncanonical Autophagy Pathway Restricts Toxoplasma gondii Growth in a Strain-Specific Manner in IFN-γ-Activated Human Cells." mBio, 6(5), e01157-15.](https://doi.org/10.1128/mBio.01157-15)
</xmp>
<br><br>
<h5 style="color: #317eac; line-height: 28px; font-size: 18pt; font-family: 'Telex',sans-serif; text-align: center"> Copyright © 2017 - <script type="text/javascript">var year = new Date(); document.write(year.getFullYear());</script> <a href="https://frickellab.com/" style="color: #2fa4e7">HRMAn team</a>,<br><a href="https://www.crick.ac.uk/" style="color: #2fa4e7">The Francis Crick Institute,<br>London</a></h5>
<script src="v/0.2/strapdown.js"></script>
</body>
</html>