The 82 significant genes taken together misclassified some of the samples, whereas 24 significant genes correctly classified the two groups of samples as shown in Fig. 2. It is evident from the above Fig. 2 that 24 significant genes neatly classified the two tumor-groups of gene expression profiles with average silhouette width value = 0.3211 as shown in Fig. 2B. In Fig. 2A, red and green points with blue circle represent the African–American check details and European–American tumors that were misclassified with 82 significant genes. In the present study, the 24 significant genes were considered as true significant

genes as they are discriminating the two tumor-groups. The scatter plot of observed t-statistic and expected t-statistic with true significant genes is shown in Fig. 3. In Fig. 3, the green points represent 58 of 82 significant genes that were present in more than 4 simulated datasets and the red points

represent 24 of 82 significant genes that were present in more than 60 simulated datasets at p-value = 0.00003. The red points with black circle represent gene symbols that are biologically related to the study and distinguish the two tumor-groups. The gene ADI1 (probeset 217761_at) is higher expressed in European-American than in African–American tumors. Similarly, the gene CNNB1 (probeset 201533_at) is higher expressed in African–American than in European–American tumors. The find more two genes, PSPH (probeset 205048_s_at) and CRYBB2 (probeset 206777_s_at) are higher expressed in African–American than in European–American tumors and these two genes are associated with race/ethnicity. All these 24 true significant genes are shown in Table 3 and discussed PAK6 in detail in gene enrichment section. It is evident from the Table 3 that there are 8 genes that are higher expressed in European–American than in African–American tumors and 16 genes are higher expressed in African–American than in European–American tumors. The twenty-four differentially expressed genes

obtained through differential expression analysis were studied further for their abundance in different gene ontology and pathways. The overabundance of a particular term was measured in terms of number of genes involved, number of genes in a particular term from the total number of differentially expressed genes (24), number of genes for a particular term in the organism’s annotation data and the total number of genes in the annotation file for Homo sapiens (54,675). Fisher’s test was used to determine the overabundance of each term in the list. Terms which are under threshold of 0.05 were taken to be the significant biological functions and pathways. It is evident from Table 4 that the functional analysis revealed clusters of terms like immune response, antigen processing, etc., showing high expression of immune response genes.