HTS QC Analysis

If at any point you need help, you can watch the movie to see these instructions being carried out.

  1. Open the primary_hts.csv file. One way to do this is to drag and drop the file onto the DecisionSite window.
  1. DecisionSite automatically displays a scatter plot. Each row in the data set is one marker on the plot. DecisionSite chooses two columns to place on the x- and y-axes; in this case it has chosen Percent Inhibition and Plate ID.
  1. Explore the scatter plot. You can click on a marker to view details about that data point in the Details on Demand window. For example, the turquoise marker in the upper left corner corresponds to Plate 3, Well B07 and represents a percent inhibition of 44.9. We can see that most of the data lies between -15 and +10 percent inhibition.

  2. Modify the scatter plot. You can change most of the attributes of the scatter plot, including the axes, the marker colors, the marker shapes, and the marker labels.
  1. If the Legend is not visible, choose View->Legend from the DecisionSite View menu.
  2. Right-click on the visualization and choose "Properties"
  3. Select the "Markers" tab in the Properties window
  4. In the "Color" section, change the color scale to "Continuous" and then change the coloring column from Plate_ID to Percent_Inhibition. Wells with high percent inhibition are now colored blue, and those with low percent inhibition are red. (Note: you can explore advanced coloring options by clicking the "Customize" button in the "Color" section).
  5. (Version 9.1 and higher only) In the "Labels" section, switch the labeling column from Well_ID to Compound_ID. When you mouse over a marker in the visualization, the label will now display the Compound ID for that marker.
  6. Close the "Properties" window by clicking the red "X" in the upper right corner.
  1. Apply Filtering. Use the Query Devices to filter out compounds that don't meet specific criteria. For example, try filtering to Percent_Inhibition > 35. Watch the number of compounds change in the lower right corner of the DecisionSite window.
  1. You should be left with around 17 compounds.
  1. View distribution of Percent Inhibition. The next few steps will help you construct a view of the distribution of percent inhibition values across the primary screen.
  1. In the Visualization menu, select New Bar Chart.
  2. Select Edit-->Reset to All to display all 4000 compounds.
  3. Change the x-axis of the bar chart to "Percent_Inhibition".
  4. Change the y-axis of the bar chart to "Count".
  1. The plot automatically groups (or "bins") the data into bars. It would be nice to see a few more bars to smooth out the distribution ...
  1. Right-click the bar chart and select "Properties".
  2. Select the "Bars" tab of the Properties window.
  3. In the lower right corner, slide the Min/Max bar to increase the number of bars in the bar chart.
  4. Close the "Properties" window by clicking the red "X" in the upper right corner.
  1. Look for QC patterns. One useful way to look for common HTS data biases such as edge effects is to average all of the plates together.
  1. In the Visualization menu, select New Pie Chart.
  2. Change the x-axis of the pie chart to "Plate_Column".
  3. Change the y-axis of the pie chart to "Plate_Row".
  1. Notice that the y-axis puts A near the bottom and H near the top. To reverse this, right-click on the Plate_Row Query Device and choose Set Property-->Reverse sorting.
  1. Right-click on the pie chart and choose "Properties".
  2. Select the "Pies" tab of the Properties window.
  3. Change the labeling to "none".
  4. Change the coloring to "Average" and the coloring column to Percent_Inhibition. Also change the sizing to "By Average" and the sizing column to Percent_Inhibition.
  5. Move the Size slider device to the right to make the markers larger.
  1. Notice how the markers near the bottom of the plate are much larger than the rest? This indicates that the average percent inhibition in those wells across the entire screen is different than the average value in the other wells. The red circle in the middle of the markers means a negative value, so these wells have - on average - more highly negative percent inhibition values than the rest of the plate. This indicates a positional bias in the data from the screening campaign.