Skip to main content
Figure 1 | BMC Bioinformatics

Figure 1

From: Mechanism-anchored profiling derived from epigenetic networks predicts outcome in acute lymphoblastic leukemia

Figure 1

Design of PGnet methods. Two sets of data are required to construct the network: (i) V ESG lists are produced by analyzing the known epigenetic "seed genes" using the pairwise standard Pearson correlation coefficient of the vsn normalized gene expression levels between ESG and all genes g n that meet the IQR filter criteria across all samples where genes g n (n = 1,..., N), (ii) V LP lists are derived by analyzing the gene's differential expression in each phenotype of interest and evaluating its significance with an adjusted t-test. The rows in both V ESG and V LP lists represent every genes in the microarray that meet the IQR filters, whereas the columns are either epigenetic seed genes in the V ESG lists or phenotypes in the V LP lists. We then use the PGnet methodology to develop a similarity vector between the epigenetic seed gene and a phenotype. We calculated the vectorial similarity between each pair of ordered expression of gene lists using a previously algorithm that we published, orderedlist (Suppl. Methods (Additional file 1)) [37]. The result is a ranked list of genes that are significantly associated based on their respective V ESGi and V LPj . We build a gene-phenotype network where relationships are similarity scores (Fig. 2). Legend: g: gene with microarray expression; ESG: epigenetic seed gene; LP: leukemia phenotype.

Back to article page