The approach described in previous sections relies on ANOVA to detect differentially methylated CpG sites and takes individual sites as a starting point for interpretation. Since ANOVA compares M values at each site independently, this strategy is robust to type I/type II probe bias.

An alternative could be to first summarize all the probes belonging to a CpG island region (i.e. island, N-shore, N-shelf, S-shore, S-shelf) and then use ANOVA to compare regions across the groups. Since the summarization will include both type I and type II probes, you may want to split the analysis in two branches and analyze type I and type II probes independently. 

Your Rating:

Thanks for voting!

Please Wait

Results:

0

rates

As the first step of the summarization to CpG island regions, transpose the β-values spreadsheet, i.e. the top level one (Transform > Create Transposed

Spreadsheet…), using the Type as the Column header. After that, right click on a column header, select Insert Annotation and choose and

USCC_CPG_ISLANDS_NAME (OK to accept). A new column will be inserted

(Figure 28).

Figure 28: β-values spreadsheet transposed and annotated by CpG islands name. The spreadsheet was sorted by column ID.

Next, right-click on the header of the new column, select Properties, and set the Type to categorical (and OK). This step is required to enable the group statistics tool (next step).

Then select Stat > Descriptive > Column Statistics… (Figure 29), check the Group by box, and move the Mean from Candidate Measure(s) to Selected Measure(s) box by using the -> button. Select OK to compute.

Figure 29: Setting the column statistics to compute means per group

The new spreadsheet (Figure 30) now has one CpG island region per row (listed in column #2, Level), samples on columns, and the values in the cells represent the mean of β-values of all the CpG probes in the region.

Figure 30: β-values summarized to CpG island regions; the spreadsheet features one region per row and samples are on columns

Note the first row, with label “– Mean”. It corresponds to all the probes that map outside of USCS CpG islands. As it is not needed for the downstream analysis, remove it by right-clicking on the row header and selecting Delete. The row will be removed permanently.

The final step is to transpose that spreadsheet back (setting the Column to 2. Level). The layout now is as follows: one sample per row with CpG island regions on columns; cell entries correspond to mean methylation status of the region (Figure 31). This spreadsheet can then be used as a starting point for ANOVA and other previously discussed procedures.

Figure 31: β-values summarized to CpG island regions; the spreadsheet features one sample per row with regions on columns