How to launch a Principal Component Analysis

How to launch and visualize a principal component analysis

HubSpot Video
  1. Create an "Analysis" within a project. Select "Statistics" and "Principal Component Analysis". 
  2. Type in any values you would like to analyze in the table or open an existing file that you have uploaded to your BioBox Library
  3. User the right hand panel to customize your PCA parameters. PCA parameters include the number of samples in the analysis, the number of components, and standardization of values.
    Screen Shot 2022-02-28 at 11.02.40 AM 
  4. Once your analysis has been computed, you will be taken to the Plot tab. User the right hand plot customization panel to specify the PCs that will be on your X and Y axis. Add sample name and PC tool tips so that when you hover over each data point you can see which sample it corresponds to.  Similar to other scatter plots on the BioBox platform, the plot is fully customizable. The plot tab allows you to swap between 2D and 3D (Sphere) scatter plots. For a comprehensive breakdown of how to customize your scatter plots, read this article here
    1. Customizing axis labels. Under the "Series" drop down, select "Axes". Use the configuration panel to specify x/y axis titles, Customize axis label size and orientation. 
    2. Adding colours to your data points.  Under the "Series drop down, select "Visual Map". Under dimension select one of the PCs that you would like to be the basis for colours applied to your plot. e.g. if PC1 is on the X axis the data points can be customized according to PC1 values. In the example data points with a minimum PC1 value of -100 and a maximum of 0 were coloured green, data points with a minimum PC1 value of 200 and a maximum of 242 were coloured red. If the data points are very close in proximity, you can add PC1 tool tips so that when hovering over the data points you see the exact values. These values are also available in your output files. 
      Screen Shot 2022-02-28 at 10.45.50 AM
    3. Labeling data points. Right click on a data point, select "Show Data Label" and under the "Labels" drop down select Sample. 
    4. Exporting your plot. Select the cloud icon in the top left hand corner of the plot and you will be able to customize the image height, width, and file type prior to export. 
    5. Under the "Files" tab you can download your PCA output files, these include PCAs and Eigenvalues. PCA output files will not be saved to your BioBox library. These files will be saved in the BioBox, if the Analysis has been deleted the PCA output files will be deleted as well.