View Source

To detect differential methylation between CpG loci in different HSCP populations, go to Analysis > Detect Differential Methylation. In the ANOVA dialog (Figure 1), select Add Factor to move the factor 2. HPSC from Experimental Factor(s) to the ANOVA Factor(s) box.

Genomics Suite Documentation > Detect differentially methylated loci > ANOVA_Setup.png

Depending on the data in the top level (i.e. parent) spreadsheet you may need to manually specify whether the data is already log transformed: you should select Yes for M-values, as they are based on logit transformation (Figure 2). By default, Partek Genomics Suite will calculate fold-change value for each contrast and, in addition to that, if you want to include the difference in methylation levels between the groups at each CpG site in the output, check the Estimate box in the Other Statistics section of the dialog (Figure 2). The setup of the contrasts has implications on downstream steps, in particular on filtering the differentially methylated loci.

Genomics Suite Documentation > Detect differentially methylated loci > Setting_ANOVA_Contrasts.png

For this exercise, we shall compare primed HPSC with suppression of OCT4 (shPOU5F1) and primed HPSC with suppression of NANOG (shNANOG) with the baseline primed cells. To start with, select the Primed group in the Candidate Level(s) box and push Add Contrast Level > to move the Primed group to Group 2 (lower box). Then Ctrl & select both shPOU5F1 and shNANOG in the Candidate Level(s) box and push Add Contrast Level > to move them to Group 1 (upper box). Then click Add Combinations and confirm that two contrast have been created as seen in Figure 3.

Genomics Suite Documentation > Detect differentially methylated loci > two_contrasts_specified.png

Push OK to confirm the contrast (and close the contrast dialog) and again to start the ANOVA calculation.

The first time you use MethylationEPIC array, the manifest file needs to be configured and the window like the one in Figure 4 will pop up. First select the second option (Chromosome is in one column and the physical location is in another column). Marker ID is on the first column (Ilmn ID), Chromosome is on the column CHR, while the Physical Position is on the MAPINFO column. Set according to Figure 4 and select Close. This enable Partek Genomics Suite to parse out probe annotation from the manifest file.

Genomics Suite Documentation > Detect differentially methylated loci > Configure_Annotation.png

The result of 1-way ANOVA is shown in Figure 5. Each row of the table represents a single CpG locus (identified by Probeset ID column). The remaining columns contain the following information:

Column 3. Gene Symbol: the gene overlapping the probe as specified in the Illumina manifest file

Column 4. p-value(HPSC): overall p-value for the specified factor (in parenthesis). A low p-value indicates that there is a difference in methylation between the levels of this attribute (i.e. study groups). The contrast p-values should then be used to evaluate individual group comparisons. If more than one factor is included in the model, p-value will be reported for each.

Next, for each contrast included in the model, a block of seven columns will be added, as follows:

Column 5. p-value(shNANOG vs. Primed): p-value for the given contrast (in parenthesis). A low p-value indicates a difference in methylation between the groups included in the contrast (here: shNANOG and Primed).

Column 6. Ratio(shNANOG vs. Primed): ratio of average methylation level in one over the other the other contrasted group (shNANOG and Primed, respectively). Ratio is reported in linear space.

Column 7. Fold Change(shNANOG vs. Primed): fold-change in one over the other contrasted group (shNANOG and Primed, respectively). Fold-change is reported in linear space.

Column 8. Fold Change(shNANOG vs. Primed) (Description): if fold-change > 1, it means hypermethylation in the first group (e.g. shNANOG up vs Primed), if fold-change < -1, it means hypomethylation in the first group (e.g. shNANOG down vs Primed), relative to the second group (Primed). This column enables quick filtering

Columns 9. & 10. Lower and upper (respectively) limits of 95% confidence interval of the fold-change

Column 11. Estimate(shNANOG vs. Primed): difference between means of two groups (i.e. shNANOG and Primed) (this column is optional and depends on the way contrasts were set up)

Columns 12. - 18. correspond to columns 5. - 11.

Columns 19.+ Statistical output

Genomics Suite Documentation > Detect differentially methylated loci > ANOVA_spreadsheet.png

Going forward, analysis of differentially methylated loci typically includes removal of the probes on X and Y chromosomes (to avoid the problems with inactivation of one X chromosome). Another common filtering is removal of probes with known single nucleotide polymorphisms (SNP) close to the interrogation site. To annotate the ANOVA spreadsheet with the information required for filtering, right-click on the Gene Symbol column, select Insert Annotation, tick-mark the CHR and SNP_DISTANCE fileds (Figure 6) and push OK. Two new columns will be appended to the spreadsheet.

Genomics Suite Documentation > Detect differentially methylated loci > Adding_Annotation.png

Section Heading

Section headings should use level 2 heading, while the content of the section should use paragraph (which is the default). You can choose the style in the first dropdown in toolbar.