- Pathway and Network Analysis 97
Figure 1. QQ plot for SAM analysis. Red points represent DE upregulated genes (when
class label 1 is use as the reference group) while green DE downregulated genes.
Ditto for DE downregulated genes:
samrSummTable$ngenes.lo
[1] 309
Note that the number of DE genesmightbe different each time when you
perform a SAM test due to the nature of the permutation procedure. We
can show the first parts of the table for DE genes using head():
head(samrSummTable$genes.up)
head(samrSummTable$genes.lo)
From the output in the screen, you learn that “CCNA|8900” is the most
significantly upregulated gene in with fold change = 8.43 and “CD52|1043”
is the most significantly downregulated gene with fold change = 0.08 in
ALL1/AF4 samples (class label 2) compared to BCR/ABL samples (class
label 1). You can go back to check the input expression matrix for these two
genes:
mergedData[“CCNA1/8900”,]
mergedData[“CD52/1043”,]
- KEGG pathway analysis
After you obtain DE genes between BCR/ABL and ALL1/AF4 samples
using SAM analysis, you may want to ask whether these DE genes can
be enriched in some molecular pathways that discern these two types of
B cell leukemia. Here, you can perform a hypergeometric test on Kyoto
Encyclopedia of Genes and Genomes (KEGG)^9 pathways using Entrez
identifiers of the DE genes. Before that, you have to obtain a mapping
between Entrez identifiers and KEGG identifiers using the Bimap object
org.Hs.egPATH in R package org.Hs.eg.db: