There are three main parts to the SGAtools pipeline. Image analysis, Normalization & scoring, and Data analysis & visualization. There are two possible start points in the pipeline shown as A, starting out with plate images and B, starting out with quantified colony sizes generated by an alternate image analysis software. It is highly recommended that input files at either start point are named using a certain convention in order to maximize functionality

The tabs on the right contain further details on each of the steps of the SGAtools pipeline.

Image analysis is the first step of the SGAtools pipeline and involves fitting a grid onto your plate images, then using the bounds of the grid to quantify the colony size using pixel intensities

Input

1. Plate images

Images of your screen in .jpg format. Images are expected to be an approximate resolution of 160 dots per inch (dpi) and should either be cropped to the size of the experimental plate (i.e. must have plate edges) or have a black background outside of the plate. Note: the higher resolution the image is, the more accurate the quantified colony size is.

---

2. Plate format

SGAtools supports 4 core screen formats

Screen format	Number of rows	Number of columns	Example
1536	32	48
768 -diagonal replicates	16	24
384	16	24
96	8	12

3. Crop option

• Automatically choose method (recommended): The program will automatically choose between the other options.
• Always autodetect plate edges: The program will locate the edges of the plate by finding the edge of the black background surrounding the plate.
• Images are already cropped to plate edges: The program will assume that the images have already been cropped to the edges of the agar plate.

3. Noise removal

Flag indicating if noise/speckles should be removed from the thresholded image prior to analysis.

4. Autorotate

Flag indicating if image should be auto-rotated prior to processing. Only select this option if image is extremely rotated. gitter is able to handle small variations in rotations (1-2 degrees) without auto-rotating.

5. Inverse

Flag indicating if input image is inverted, meaning colonies are darker compared to their background.

Type	Example image
Bright colonies on darker background.
Dark colonies on lighter background

Output

Input

1. Plate files

The main input to normalization & scoring is the quantified colony sizes in a space-delimited file with 3 or more columns as follows:

1. Row
2. Column
3. Quantified colony size

# This is a commented line and is ignored
# rows	columns	size	circularity: 
1	1	2205	0.981750  
1	2	1734	1.065585  
1	3	1996	1.057621  
1	4	1704	1.032656  
1	5	1755	1.109302  
...
					

Note: In the example above, the circularity column and any column that follows it are ignored

---

2. Plate layout file

This file allows a mapping from row and columns to a gene name. If an array layout file is not selected the array column in results will contain numbers, such that each number represents a group of replicates.
There are two ways of selecting an array layout file:

Selecting a predefined plate layout

You can select between commonly used arrays from the drop down menu. The number of the plate detected from the file name (e.g. plate 4 for user_ctrl_YOR101W_4_130301.jpg) is matched to the corresponding plate in the array, and the output is populated with the names of the strains in the plate.

Uploading a custom plate layout file

Select the "Upload custom" option for plate definition, and upload your file. The format for the plate definition file is three tab-delimited columns:

1. Column
2. Row
3. Gene name

A name is assigned to each replicate group. Below an example shows the head section of an array layout file along with a graphical representation.

# This is a comment line and is ignored		    			
# c	r	Gene
1	1	HIS3
1	2	HIS3
1	3	HIS3
2	1	HIS3
2	2	RKM3
2	3	YPK3
3	1	HIS3
3	2	TPS1
3	3	GAL1
...
    						

---

3. Replicates

This is the number of replicates in the experiment.
A value of 4 indicates replicas are in quadruples
A value of 1 indicates there is only one replicate

---

4. Linkage correction

This step ignores interactions between genes within a specific proximity to one another on the same chromosome from the analysis as it is considered an artifact. The proximity is provided in kilobases (KB) as the linkage cutoff

---

5. Score results

If this option is selected, a score will be computed for non-control plates with a corresponding control plate of the same array plate id. There are two ways SGAtools can score screens:

1. 
2. 

Such that Cij represents the fitness of the double mutant, Ci the single mutant fitness of the query, Cj the single mutant fitness of the array

Output

This is the final step of the SGAtools pipeline and involves visualizing your processed data. There are two visualizations available:

1. Heatmap: If you analyzed your images with SGAtools, your plate will appear adjacent to a heatmap of your processed data. Hover over the heatmap for details of colonies
   

---

2. Histogram: An interactive histogram of the data is shown. Click and drag the histogram to create a window of data you would like to inspect
   

There are 3 types of processed data you can visualize:

1. Raw colony sizes: The colony sizes as quantified by the image analysis software
2. Normalized colony sizes: The colony sizes normalized by SGAtools
3. Score: Colony score against control plate (this is only available if the plate is not a control and sufficient data was provided to score)

Select a specific plate from the drop down and the type of data you would like to visualize. You can also adjust the range of values to be displayed.