Procedure description

Quick Start

This section describes in brief the main steps of microarray image processing using Spotfinder.

  1. Check program settings. On Menu Bar go to Settings->General Settings. Make sure that the Channel Number is 2 and Scale Factor is 2. For the first run uncheck the check boxes Use RI plot and Use Full Table for MEV data: we are not sure you computer is advanced enough to digest the full data set. More details...
  2. Load Images. Select both images at once by holding keyboard Ctrl- key to click on one file then on another one.
  3. Load grid from file. Alternatively compose grid from scratch. More details...
  4. You should be switched to Analysis Page. If not go to it. Move manually whole set of grids to overlay with the image. Place the top left grid in correct position using All Grids Menu, ignore at this moment other grids. Place top left cell in correct position over top left spot on first (top left) grid. Expand or shrink ALL grids in both dimensions. If needed, change appropriately the cell size for ALL grids. More details ...
  5. Align grid automatically or manually. To make automatic alignment enter mask size on Gridding and Processing Dialog and run Adjust All .More details...
    To align grid manually: on Analysis Page use Move All from All Grids Menu at first to move all grid set in proper location.
    If it is needed use Move for every grid separately. More details...
    Store aligned grid template in SFG file.
  6. Set Processing parameters in Gridding and Processing dialog: set Segmentation Algorithm to Otsu. Leave deafault settings for Minimum and Maximum spot size on this dialog. Click Process All button. More details...
  7. Check the alignment on QC View Page. Compare the the size of green area on QC View. The grid with green area larger than that of the rest grids on array indicates a possible grid misalingment. Go to Analysis Page to check this and realign if it is necessary.
  8. Create MEV file. On Menu Bar go to Data, select Save Data to MEV file.

Procedure steps in details

To adjust grid size:

  1. Move the whole grid set to align top left spot with top left cell of first grid. To do it click right mouse button when mouse pointer does not overlay with any grid. All Grids menu will be activated. Choose Move All from the All Grids menu and move mouse sligthly or use keyboard arrows keys to move all grids together to necessary position. When it is done terminate Move mode by pressing End keys on keyboard or click left mouse button.
  2. If number of rows and columns is set correctly but grids do not fit with the image spots you might need expand or shrink grid. Invoke All Grids menu and choose Expand All command. By using keyboard arrows keys user can expand or shrink all grids simultaneously in any dimension. Both expansion and shrinking are performed while keeping left and top edges of each grid fixed. Only the right and bottom edges of each grid are moved.
  3. If it is neccessary adjust cell size using command Cell Size All from All Grids Menu. Use keyboard arrow keys to increase or decrease cell size in vertical and horizontal dimension. One arrow key click corresponds to cell size increase/ decrease by one pixel. Terminate Cell Size All mode when finish by clicking End on keyboard or left mouse button. When increase cell size try not to touch the spots in neighboring cells by the adjusting cell box. Unless the adjusting cell box does not reach the adjacent cells it is safe and desirable to incease the area around spot for local background calculation.
  4. User can rotate grids if needed. Activate this command mode Rotate All from All Grids Menu and use Up and Down arrows keys on keyboard to rotate all grids together. Rotation of each grid is performed around its top left corner. Again, at this stage use only top leftmost grid as indicator of alignment. The rest of the grids are expected to have the same angular offset due to the parallel orientation of pins during the slide printing.
  5. User can repeate steps 1-4 in any sequence any times to reach better grid size adjustment. Make sure at this point to use only first grid (top left) as indicator of proper grid size .

To adjust grid locations on image:

  1. By this point first grid (top left) has to be aligned perfectly in size, orientation and location. The rest of the grids should have proper size and orientation, but probaly not location. They may have some misalignment in vertical and horizontal position because the pins in reality are not exactly parallel to each other in the robot printing head. When grid size and orientation are set user can start to adjust location for all grids. To make automatic alignment enter mask size on Gridding and Processing dialog and push Adjust All button or choose automatic command Adjust All from All Grids menu. The mask size should be larger than the typical spot size on array but smaller than the cell size or spacing. It applies adjustment algorithm to each grid one by one and find the best location for each grid on image. This procedure can be applied as many times as it needed.
  2. User also can apply this automatic procedure individually to each grid using Grid Menu command Adjust. To do it user needs to move mouse pointer on top of the selected grid, this grid becomes colored in magenta. Then pressing right mouse button user invokes Grid Menu and selects Adjust from the menu list. This procedure can be repeated as many times as needed.
  3. If automatic Adjust procedure does not provide satisfactory alignment user can manually move selected grid. Select a grid by placing mouse pointer on top of it and activate Move mode from Grid Menu by right clicking mouse. By moving mouse or applying the keyboard arrow keys user can move selected grid to proper location. When done terminate Move mode by pressing keyboard End or left mouse button.

To process spots in cells:

  1. Before to start of grid processing user must enter experimentally defined parameters: estimates of Spot Size, which is the typical spot diameter in pixels. Depending on selected algorithm for spotfinding user must set one or two values for Spot Size . Currently three spotfinding methods are implemented in the program.
    Histogram thersholding method was historically the first implemented in Spotfinder. It requres only one value to enter: the estmate of Spot Size in pixels. This value can be estimated by zooming image in Analysis page and placing mouse pointer on left and right edges of typical spot (alternatevely, top and bottom edges of spot). By reading the X (or Y) coordinates in image pixels on Status Bar and calculating difference the spot size estimate can be done. It is safer to use slightly overestimated then underestimated value for Spot Size. For this reason if it exists a noticable difference in spot sizes across the image it is better to use largest value for Spot size.
    Otsu thresholding method is default and requires user to set two values for spot size: minimum and maximum. These values specify the range, which program uses to search for spot boundaries. Default settings for Otsu algorithm cover almost all possible spot sizes observable in real microarray images. However, we recommend to narrow it to the expected real spot size range. For Histogram thresholding set Spot Size in one entry boxes on Gridding and Processing dialog. The rule of thumb: the entry should be slightly larger than the biggest spots on array.
    Fixed Circle segmentaion method is the most simple of all. It takes as a spot the area inside the circle centered in the cell. The only parameter user must set is a circle diameter in pixels.
  2. Press the Process All button, or go to Process All in All Grids Menu. Spotfinder runs through all grids one by one. Each cell in the grid is analyzed to detect spot boundaries using the specified Spot Size and data from both channels A and B. After removing the saturated pixels from analysis, intensities for signal and background are calculated. Spotfinder holds internally the all raw data after processing. The next step is post processing data maipulation, which is performed automatically in a row with processing. In post-processing steps background subtraction is performed and background corrected integrated inetnsities or medians are reported by default. Also the default QC filter procedure is applied and spots flags are set. More about QC filter and quality controls.... User can reset default setting for post-processing any time and update reported data set on RI Plot Page and Data Page. For more details of post-processing data manipulation go to Post-Processing Dialog
  3. It is possible to process each grid individually by starting Process from Grid Menu of selected grid.
  4. Check box Contour to visualize the spots boundaries colored in accordance with flags set by QC Filter. Good spots are colored in magenta. Spot's contour is drawn in green if QC filter defined spot as bad at least in one channel.