Table 1.
RNA sequencing statistics.
Table 2.
RNA sequencing statistics by sample.
Fig 1.
(A) Multidimensional scale plot of PDAC tumor (red bar) and normal (blue bar) RNA-seq data. (B) Volcano plot summarizing the differential gene expression analysis, showing log2 fold change vs. log10 of the p-value (BH adjusted). Red indicates genes that were > = 1.5 log2 fold change and had an adjusted p-value < = 0.05. (C) The heatmap summarizes the unsupervised clustering and differential gene expression analysis comparing the normal samples to the tumor samples. Dendrograms show the unsupervised hierarchical clustering of samples (top) and genes (left side). A larger version of this heatmap with samples and genes labeled is provided in the supplementary results (S1 Fig in S1 Appendix).
Table 3.
Select gene set enrichment analysis results.
Cluster profiler was used to probe genesets at the molecular signature. Database using gene differentially expressed between tumor and normal samples (Yu et al 2012, Subramanian et al 2005, Liberzon et al 2015).
Fig 2.
Heatmaps of geneset genes that are differentially expressed between tumor (red bar) and normal (blue bar) PDAC samples.
(A) Hallmark Complement Geneset, the C2 gene is indicated by (*). (B) Hallmark Epithelial Mesenchymal Transition Geneset. (C) Hallmark Inflammation Response Geneset. (Ref = PMID 26771021).
Fig 3.
Boxplots of log2 gene counts for C2 (left) and C3 (right) in normal samples (blue) and tumor samples (red).
The difference in C2 expression between normal (mean = 3.48; SD = 1.16) and tumor samples (mean = 16.01; SD 0.91) was significant (t (13) = -2.53 p < 0.0001). The difference in C3 expression between normal (mean = 9.92; SD = 1.31) and tumor samples (mean = 11.51; SD 0.85) was significant (t (13) = 5.14 p = 0.0002).
Fig 4.
CNET plot showing gene overlaps in enriched genesets.
Geneset enrichment analysis of genes differentially expressed between tumor and normal samples indicated enrichment for sets of genes up (olive) or down (orange) regulated in pancreatic cancer, the complement cascade (green), epithelial mesenchymal transition (blue), and inflammatory response (purple). Genes at the end spokes indicate differential expressed in our dataset that were enriched in each geneset, color of dots indicated whether the gene was up (red tones) or down (blue tones) regulated in tumors relative to normal samples [17, 21, 35].
Fig 5.
CNET plot showing gene overlaps in enriched genesets.
Geneset enrichment analysis of genes differentially expressed between tumor and normal samples indicated enrichment for sets of genes involved in KRAS signaling (purple), inflammatory response (turquoise), complement cascade (red), and epithelial mesenchymal transition (green). Black dashed circle indicates PLAUR, which connects to all four genesets. Genes at the end spokes indicate differential expressed in our dataset that were enriched in each geneset, color of dots indicated whether the gene was up (red tones) or down (blue tones) regulated in tumors relative to normal samples [17, 21, 35].
Fig 6.
Boxplots of log2 gene counts for CLEC7A (DECTIN-1 (left)) and LGALS3 (Galectin-3 (right)) in normal samples (blue) and tumor samples (red).
The difference in CLEC7A expression between normal (mean = 72.85; SD = 133.99) and tumor samples (mean = 288.16; SD = 142.84) was significant (t (13) = 3.86 p = 0.002). The difference in LGALS3 expression between normal (mean = 43.80; SD = 62.85) and tumor samples (mean = 358.13; SD 206.68) was significant (t (13) = 6.43 p < 0.0001).