Figure 1.
Gene expression profiles hierarchical cluster analysis based on Spearman correlation.
Spearman correlation of gene expression profiles is used to define the gene expression profiles similarity of 21 different tissues/cells. A hierarchical cluster analysis with correlation information shows 2 clusters: 12 normal tissues and 9 cancer cell lines.
Table 1.
RNA-Seq sample source.
Figure 2.
(A) Comparison of normal and cancer HK genes definition.
HK genes are defined separately from two physiological groups: 12 normal tissues and 9 cancer cell lines. (B) Different HK gene types functional enrichment. “Cancer” means cancer HK genes, abbreviated as suffix “C” follows function term illustration; “Cancer-associated” means specific HK genes in cancer condition, abbreviated as suffix “CA” follows function term illustration; “Shared” means overlapped HK genes in normal and cancer conditions, abbreviated as suffix “S” follows function term illustration; “Normal-unique” means specific HK genes in normal condition, abbreviated as suffix “NU” follows function term illustration; “Normal” means normal HK genes, abbreviated as suffix “N” follows function term illustration.
Table 2.
Functional enrichment analysis of normal and cancer HK genes.
Table 3.
Low and high gene expression thresholds calculated by the improved K-means algorithm.
Figure 3.
(A) Shared HK genes expression variation distribution in normal and cancer condition. (B) Gene expression variation adjustment in shared HK genes between normal and cancer condition.
There are three gene expression variation statuses, Constant, abbreviated as suffix “C” in (B), and Moderate variable, abbreviated as Moderate in (A) and suffix “M” in (B), and Variable, abbreviated as suffix “V” in (B).
Figure 4.
MDAD plots of shared HK genes.
MD<0 means the gene expression span width in cancer condition is larger than that in normal condition, and AD<0 means the gene expression relative average level in cancer condition is higher than that in normal condition. According to shared-normal HK genes expression variation statuses, shared HK genes are divided into three subtypes, constant, moderate variable, and variable expressed shared HK genes. Paired Wilcoxon signed rank test is used here to measure gene expression regulation and gene expression variation status regulation. (A) All shared HK genes. (B) Shared constant expressed HK genes. (C) Shared moderate variable expressed HK genes. (D) Shared variable expressed HK genes.
Table 4.
Comparison of paired Wilcoxon signed-rank test results of MDAD values for shared HK genes in normal and cancer condition.
Figure 5.
MaxR and minR value distributions of shared HK genes.
Up y-axis denotes maxR with range [0, 3], and down y-axis denotes minR with range [0, 3]. For amplifying the figure, we set the ratio value as 3.00 if a ratio value is larger than 2.50. As to the inner insert graph, the blue curve shows accumulated maxR; and the green curve shows accumulated minR. Both correspond to left y-axis signifying accumulated gene count. Right y-axis denotes individual gene count (shown as Gene Count Ratio), which corresponds to a red maxR distribution curve and a cyan minR distribution curve. We quantify the proportion of genes for which cancer cell modulate gene expression level to be higher than that in normal physiological status. (A) All shared HK genes. (B) Shared constant expressed HK genes. (C) Shared moderate variable expressed HK genes. (D) Shared variable expressed HK genes.
Table 5.
Accumulated shared HK genes ratio when minR and maxR≤1.
Figure 6.
Six types of HK genes Coefficient of Variation (CV) values distributions.
The up and down bars signify Q1 (one quarter) and Q3 (three quarters) of CV values, marked as constant and variable expression threshold values. “Normal” CV values distribution for normal HK genes; “Normal-unique” is CV values distribution calculated from specific HK genes in normal condition; “Shared-normal” is CV values distribution in 9 cancer cell lines calculated from overlapped HK genes in normal and cancer conditions; “Shared-cancer” is CV values distribution in 12 normal tissues calculated from overlapped HK genes in normal and cancer conditions; “Cancer-associated” is CV values distribution calculated from specific HK genes in cancer condition; “Cancer” is CV values distribution for cancer HK genes.
Figure 7.
MDAD plots of cancer-associated HK genes.
MD<0 means the gene expression span width in cancer condition is larger than that in normal condition, and AD<0 means the gene expression relative average level in cancer condition is higher than that in normal condition. According to cancer-associated HK genes expression variation statuses, cancer-associated HK genes are divided into three subtypes, constant, moderate variable, and variable expressed cancer-associated HK genes. Paired Wilcoxon signed-rank test is used here to measure gene expression regulation and gene expression variation status regulation in cancer. (A) All cancer-associated HK genes. (B) Cancer-associated constant expressed HK genes. (C) Cancer-associated moderate variable expressed HK genes. (D) Cancer-associated variable expressed HK genes.
Figure 8.
MaxR and minR value distributions of cancer-associated HK genes.
Up y-axis denotes maxR with range [0, 3], and down y-axis denotes minR with range [0, 3]. For amplifying the figure, we set the ratio value as 3.00 if a ratio value is larger than 2.50. As to the inner insert graph, the blue curve shows accumulated maxR; and the green curve shows accumulated minR. Both correspond to left y-axis signifying accumulated gene count. Right y-axis denotes individual gene count (shown as Gene Count), which corresponds to a red maxR distribution curve and a cyan minR distribution curve. We quantify the proportion of genes for which cancer cell modulate gene expression level to be higher than that in normal physiological status. (A) All cancer-associated HK genes. (B) Cancer-associated constant expressed HK genes. (C) Cancer-associated moderate variable expressed HK genes. (D) Cancer-associated variable expressed HK genes.
Table 6.
Comparison of paired Wilcoxon signed-rank test results of MDAD values for cancer-associated HK genes in normal and cancer condition.
Table 7.
Accumulated cancer-associated HK genes ratio when minR and maxR≤1.
Figure 9.
Gene structure bias among different HK gene types.
Six types of gene structure are focused here, including: gene size, CDS length, number of exons, number of minimal introns, number of large introns, and GC content. We compare five types of HK genes to the background of genome (G, black line), including: cancer HK genes (C, red line), normal HK genes (N, blue line), cancer-associated HK genes (CA, purple line), normal-unique HK genes (NU, cyan line), and shared HK genes (S, green line). This reflects gene structure and sequence content affect transcription in different physiological conditions. (A) GC content bias among different HK gene types. (B) Gene size bias among different HK gene types. (C) Six types of gene structure bias in cancer-associated, normal-unique and shared HK genes, compared to the background of genome. (D) Six types of gene structure bias among five different HK gene types compared to the background of genome.