‹‹ Back to SVS Home

Visualizing Copy Number Analysis Results

10.7 Visualizing Copy Number Analysis Results

There are several ways to visualize copy number analysis results. It all depends on what results you want to visualize. Below are several different ways to visualize copy number data. These ways are not exhaustive, but are indicative of the typical ways CNV data is viewed.

Log2 Ratios

If the quantile-normalized intensities of the log2 ratio data are to be plotted for visual inspection for the presence of CNVs, this can be done from the log2 ratio spreadsheet with samples columnwise. In order to plot each sample’s log2 ratio intensities versus genomic position, the spreadsheet with intensities as columns needs to be transposed so the marker names are row labels and the sample names are column name headers. The transposed spreadsheet can also be attained by checking the “Samples Columnwise” spreadsheet during the CEL import process. See Affymetrix Files for more information about this process. See Transposing Spreadsheets for instructions on how to transpose a spreadsheet. Transposing can take some time for large datasets, so if it is known in advance only a subset of the data is to be plotted, inactivate all rows not needing be transposed. The transposing operation only transposes active rows and columns.

From the transposed spreadsheet, select Plot > Numeric Values. This will bring up a Numeric Value Plot Parameters window. Select all or a subset of samples to plot, and set the plotting parameters as desired. See Genome Browser for more information.

CNV Segment Mean Covariates

If the segment mean covariates of the log2 ratio data is to be plotted for visually inspecting CNVs, this can be done from the CNV Segment Covariates spreadsheet (preferably the spreadsheet with a column for every marker). This spreadsheet needs to be transposed so that the marker names are row labels and the sample names are column name headers in order to plot each sample’s segment mean versus genomic position. See Transposing Spreadsheets for instructions on how to transpose a spreadsheet. Transposing can take some time for large datasets, so if it is known in advance that only a subset of the data is to be plotted, inactivate all rows that do not need to be transposed. The transposing operation only transposes active rows and columns.

From the transposed spreadsheet, select Plot > Numeric Values. This will bring up a Numeric Value Plot Parameters dialog. Select all or a subset of samples to plot, and set the plotting parameters as desired. See Genome Browser for more information.

A heat map of segmentation covariates is a good tool for visually detecting interesting copy number regions. This can be done with either the segment mean covariates or the discretized covariates. It is useful to first sort the samples by case/control status in order to easily inspect the top and bottom halves of the heat map. To do this, merge the CNV Segment Covariates spreadsheet with a phenotype spreadsheet containing case/control status for the samples, then sort on the column containing case/control status. From this spreadsheet, select Plot > Heat MapNo prob.

CNV Segment Means Histogram

Plotting the histogram of segment means can be useful in visually identifying thresholds between copy number states of loss, neutral and gain for a dataset. To plot the histogram, select Plot > Histograms from the CNV Segment List Spreadsheet, and from the Histogram Parameters dialog select the “Segment Mean” column. Additional parameters can be changed from their defaults. See Histograms for more information.

CNV Segment Counts Histogram

Plotting the histogram of segment counts can also be useful in visually identifying noisy samples with a large number of segments. To plot the histogram, select Plot > Histograms from the CNV Segment Counts Spreadsheet, and from the Histogram Parameters dialog select the “Segment Counts” column. Additional parameters can be changed from their defaults. See Histograms for more information.

Log2 Ratios and CNV Segments Together

It can be a useful visual tool to plot CNV segments on top of original log2 ratio data. To do this, make sure a marker map is applied to both the original log2 ratio transposed spreadsheet and the CNV segments transposed spreadsheet.

From the original log2 ratio transposed spreadsheet, plot the sample or samples of interest by going to Plot > Numeric Values.

Next, select the first log2 ratio graph node and click on the Add Item tab in the graph control interface. Click on the spreadsheet selection menu and select Select Spreadsheet.... In the spreadsheet selection dialog select the CNV segments transposed spreadsheet. Once the columns have been loaded into the plot selection list box, select the same sample for the graph that had the log2 ratio values plotted.

In general, it is usually desired to have the CNV segments be plotted as a step line, and to leave the log2 ratio values as points on the plot. To change the options for the segment mean covariate log2 ratios, select the appropriate plot item under Graph 1 in the Graph Control Interface pane on the left-hand side of the Plot Viewer. On the Item tab change the line style to Steps, and the line weight to 2. The line can be brought to the front of the viewer by clicking on the name of the plot item and dragging it above the first plot item. The plot items can be renamed by right-clicking on the name and selecting “Rename”.

More graphs can be added for additional samples by selecting “Add New Graphs” in the Visualization box and selecting both occurrences of another sample. The above procedure can be repeated for as many graphs as desired. See Visualizing Data for more information.