Gene expression (RPKM), log-fold change relative to control liver, and significance were calculated using the DESeq software package. Individual gene transcripts were considered de-regulated if the false discovery rate adjusted p-value (q-value) was less than 0.05.
For the samples Tumor (T), Regression Day 3 (R3), and Recurrent Tumor (RT), the total number of significantly de-regulated transcripts exceeded 5000, which is the maximum capacity for pathway analysis using Ingenuity Pathway Analysis. In order to decrease the number of genes included in the pathway analysis, a subset of these genes was used according to the following criteria:
Pathway significance is indicated with log2-transformed p-values, calculcated by IPA using Fisher Exact tests. P-values were false-discovery corrected using the Benjamini-Hochberg method, and were considered to be significantly de-regulated when the corrected p-value was < 0.01.
Activation Z-scores are pathway indicators calculated by IPA which reflect the predicted activation or inhibition of a pathway. Z-scores are calculated using a repository of information from the literature about how each gene in a pathway is predicted to contribute to overall pathway activation or inhibition. Positive values indicate predicted pathway activation, and negative values indicate predicted pathway inhibition.
Not all pathways have sufficient evidence in the literature about how each gene contributes to overall pathway activity, so Z-scores are not calculated for all pathways. Pathways which lack evidence to predict activation or inhibition are designated by "N/A."
Ratios are calculated by IPA as the number of significantly de-regulated genes compared to the total number of genes in the pathway. Since IPA has a limit on the number of genes that can be simultaneously analyzed for pathway analysis (5000), the number of significant genes which IPA "sees" during its pathway analysis may differ from the true number of significant gene differences, particularly in the samples Tumor, Regression Day 3, and Recurrent Tumor. For these samples, only a subset of genes was able to be sent for pathway analysis, so the true ratio of dysregulated genes may be higher.