Skip to content

Differential Expression Analysis

After generating and filtering out lower expressing genes from the median ratio normalized expressions data, it is time to perform differential expression analysis to see if there are genes or transcripts (trancripts will be used here) that are statistically significantly up or down-regulated between biological conditions (in this case tumor versus normal).

Refer to the "Differential Analysis" section of the Partek Flow documentations to learn about the options for performing this task but in this example, DESeq2, which is Partek's implementation of the DESeq2 R package will be used. The video below shows the steps for completing differential expression analysis, constructing a volcano plot, and filtering out the differential expression results for use with over representation analysis.

After differential expression analysis is completed, a data node is generated. Click on it to review the results table. On the left, there is a panel where users can filter the differential expression results based on criteria such as false discovery rate (FDR) and fold change.

Under the "View" column of the differential expression results table, researchers can obtain a dot plot of the expression for the corresponding transcript or gene across all samples as well as a summary of the differential analysis results for that particular transcript or gene.

Other columns include

  • Gene ID
  • Transcript ID
  • Gene name
  • Transcript name
  • P-value
  • False discovery rate
  • Ratio: expression ratio obtained by dividing the mean expression of a gene in one condition by the mean expression of a gene in another (ie. numerator/denomintor, in this case tumor/normal as set during configuration of differential expression analysis).
  • Fold change: this is equal to ratios that are greater 1; when ratio is less than 1, then Fold change is equal to -1/ratio.
  • LSMean(tumor): mean expression for a given gene/transcript in the tumor samples
  • LSMean(normal): mean expression for a given gene/transcript in the normal samples

Click on the "volcano" on top of the differential expression results table to view a volcano plot. In RNA sequencing, the volcano plot displays log2 of fold change on the horizontal axis and -log10 of the p-value on the vertical axis. The plot below was filtered such that the points labeled correspond to genes whose log2 fold change value are less than -10 or greater than 10 while p-value is between 0 and 0.001.