The overall goal of this procedure is to use an image analysis platform to rapidly analyze and explore fluorescent microscopy images. This is accomplished by first seeding cells into a plate, adding perturbations to each well and incubating the plate. Images are then acquired using an automated fluorescence microscope, and the images are semi automatically loaded onto the pheno ripper image analysis platform using the pheno loader tool.
After setting the analysis parameters, pheno ripper automatically identifies recurrent block types in the images and generates phenotypic profiles for each image that can be explored using the pheno browser. Ultimately, the final cluster gram shows relationships between different conditions and image features, allowing for the comparison of complex yet subtle phenotypes generated by the high throughput imaging screen. The main advantage of this technique over existing methods that require automated identification of individual cells is that it is much faster and easier to use.
This method can help answer key questions in the biomedical field, such as how different drugs affect cellular processes. This procedure begins with cell culture and cell fixing as described in the text protocol stain the fixed cells using the specific primary antibodies, fluorescently labeled secondary antibodies and other fluorescent stains that can be found in the text protocol. After cell staining, acquire images using an epi fluorescence microscope.
In this example, we used 20 x magnification with one by one camera.Benning. We acquired 16 images for each well for a total of 480 images. To analyze images, first, download and install pheno ripper from pheno ripper.
org and launch pheno ripper. To begin analysis of images to semi automatically load images and associated metadata, click file structure to access pheno loader in pheno loader. Select the directory containing the image data set and specify the file extension of the images if required.
Filter files that are not used for analysis. Click on define markers to specify the biomarkers used in the experiment and associate them with the images. Specifically click on an image file and then select a portion of the file path that uniquely identifies the biomarkers.
Finally, enter a name for each biomarker. The next step is to click on define metadata. To associate information with each image, click on an image file and then select a portion of the file path identifying image information.
A window asking for a group name will pop up. Enter a name. Members of the group extracted from the file name will be shown on the group table.
This information can be used during visualization of the analysis results. To add additional information not present in the file name, click on the add additional information button, then select Add metadata group, and enter the additional group name. Followed by each value for the predefined group, there is no limit to the number of additional groups that can be added.
Finally, click on create metadata file to generate a pheno ripper readable annotation of the data, select the metadata field that defines replicates in the experiment to avoid redundant sampling. For a data set that contains replicates, grouping them together can improve performance. Next, click on set parameters to define the parameters required to process the data set.
The parameters to define are on the left panel. The right panel displays emerge. Sample of the data set.
Click on the left or right. Arrows to switch among different images is choose an appropriate threshold value to roughly identify the cellular portions of the images. A threshold is pre-calculated by pheno ripper, but may be adjusted to improve highlighting of cellular regions using the scroll bar available on top of the image.
If a chosen threshold does not work across all images, lower the threshold so as to include non cellular regions rather than dropping cellular regions. Next, choose an appropriate block size measured in pixels to subdivide the image into a grid of blocks. Adjust the block size to obtain an average of 20 to 30 blocks per cell to superpose the grid over the image.
Click on the checkbox in front of the block size field. If the auto scaled images do not display markers with the desired relative intensities or if markers need to be dropped from the analysis, use the adjust channel intensity option to identify cellular regions on a subset of markers. Select the channels used for threshold option default.
Choices of other parameters are usually adequate, but if required, adjust them. Using the advanced analysis options, run the analysis by pressing pheno rip on the main application panel. Pheno ripper will automatically identify recurrent block types in the images.
Phenotypic profiles will be generated for each image based on the fractions of the different block types in it. Explore the relationships between the phenotypic profiles of the images using the pheno browser. The pheno browser interface consists of four panels.
The top left panel is a graphical representation of the relative similarity between different images and experimental conditions. The two panels on the right display sample images of the selected conditions. The final panel compares the phenotypic profiles of the selected conditions.
First turn off grouping from the data processing menu to examine individual images. Next, click through the outlier points in the scatter plot and discard low quality images. Decide which metadata field defines a basic unit of comparison.
For example, to compare drugs group together images with the same value of the metadata field drug. Each point now represents a single drug using the data display menu, color or label points based on available metadata to aid interpretability. The distance between points reflects the relative phenotypic similarity of the corresponding experimental conditions.
Closer points suggest more similarity. The three dimensional plot is rotatable by selecting the rotatable checkbox. Drag the mouse over the graph while holding down the left mouse button.
Select two different points to compare them directly. The image panels will display sample images for each and the bar chart panel will show the block types that best distinguish the two points. Launch the cluster gram from the cluster gram menu to provide a more global view of the relationship between block types and experimental conditions by grouping both block types and conditions.
This representation highlights the cellular phenotypes that best distinguish groups of conditions. Pheno ripper was used to analyze hela cells that were fluorescently stained for DNA Acton and alpha tubulin and imaged after treatment with histone deacetylase, inhibiting microtubule targeting and DNA damaging drugs. The foreground threshold of these images was set to 5%and the block size to 20 pixels analysis of this dataset took approximately 10 minutes on a standard desktop computer.
The pheno browser interface was used to explore the results. The relative phenotypic similarities between images are visualized using multidimensional scaling. Each point represents a single image colored by its drug's mechanism of action.
This plot shows that phenotypic profiles can be used to group drugs by their mechanism of action. Imaging artifacts such as pore staining and empty frames were easy to identify since they stood out as clear outliers. A cluster gram of the data relates this grouping to the phenotypic profiles here.
Each row represents a single drug and each column a block type. The gray scale values reflect the phenotypic profiles. A lighter value indicates a higher fraction of that block type hierarchical clustering largely separates the different drugs by drug class.
Once mastered, this technique can be done in 10 minutes if it is performed properly. Following this procedure, other methods such as IRA profiling can be performed. This will help answer additional questions such as how different genes affect similar pathways.
