Transposable Elements-Derived MicroRNA Expression Patterns in TCGA Dataset for 10 Species

Abstract

MicroRNAs (miRNAs) are a class of non-coding RNAs that act as regulators of disease. An evolutionary approach to the disease could reveal factors such as diagnosis, treatment, and prognosis prediction. The expression patterns of transposable element (TEs)-derived miRNAs could help elucidate diseases, and their evolutionary patterns are also valuable. The 34 miRNAs were compared in terms of stage survival and tumor status in 33 carcinomas from TCGA. Expression levels were compared using a t-test and presented as differentially expressed miRNAs (DEMs). For DEMs showing statistically specific expression patterns for 3 conditions (normal and cancer, early and advanced stage, and survival), interactions with related genes in 10 species, including humans, were compared. The enrichment term was discovered for the gene-miRNA interactions. In 18 out of the 33 carcinomas, at least one miRNA was retrieved with P < .05 and |fold change| >.05. A total of 128 DEMs for the 9 miRNAs were identified. Based on the TargetScan database, interactions between miRNAs and genes in 10 species, including humans, were confirmed. The evolutionarily conserved miR-130a was observed in all 10 species, whereas miR-151a was only observed in humans. GO terms of related genes were selected for the miRNAs commonly found in each species. Evolutionary analysis of TE-derived disease-associated miRNAs was performed, and the evolutionarily conserved miR-130a-related carcinomas included renal and thyroid cancers. Human and rhesus monkey-specific miR-625 is associated with various carcinomas.

Keywords: MicroRNA; The Cancer Genome Atlas (TCGA); transposable element.

Figure 1.

Flowchart of this study. The expression level of TEs-related miRNA by cancer type was confirmed based on small RNA-seq patient information through 33 cancer types of TCGA. The expression pattern confirmed the stage of cancer, cell survival, and the expression level between normal and cancer tissues. A total of 33 disease-related TE-derived microRNAs (excluded has-miR-546b) were analyzed. Differentially expressed miRNAs (DEMs) were compared and analyzed by distinguishing them into early states (stage 1, stage 2) and high states (stage 3, stage 4) according to cancer progression. In addition, comparative analysis of carcinoma and normal tissue, and finally dead and alive samples were also comparatively analyzed.

Figure 2.

Transposable-elements derived miRNAs and their relevance to diseases. Black cells indicate that an association between miRNA and disease has been reported. A total of 34 miRNAs and their associations with 24 diseases were illustrated on a heatmap. For each row and column, k-means clustering algorithm-based classification was performed, and the rows and columns that were close to each other were grouped. Abbreviations: ACC, acinar cell carcinoma; AD, Alzheimer’s diseases; BLCA, bladder carcinoma; BRCA, breast carcinoma; CEC, cervical carcinoma; COC, colorectal carcinoma; Dem, dementia; EC, endocrine carcinoma; GBM, glioblastoma; GC, gastric carcinoma; HPV, human papillomavirus; KIC, kidney carcinoma; L, leukemia; LIC, liver carcinoma; LUC, lung carcinoma; NBM, neuroblastoma; OV, ovarian carcinoma; PAC, pancreatic carcinoma; PC, preeclampsia; PCC, pheochromocytoma; PGG, paraganglioma; PRC, prostate carcinoma; SKCM, melanoma; UCEC, uterine corpus endometrial carcinoma.

Figure 3.

Visualization of the expression patterns according to 3 classification criteria for 9 miRNAs. Differentially expressed miRNAs (DMEs) between 3 classification criteria (stages, survival, and compared to normal) were represented as heatmaps for a total of 9 miRNAs (A). The expression of each cancer type was compared according to normal and cancer tissues, early (I, II) and advanced (III, IV) stages, and survival. Compared to normal, overexpression in tumor, high stage, and dead samples was indicated in red, otherwise in green. The degree and statistical significance of miRNAs with different expressions according to the 3 classification criteria in a total of 18 carcinomas were obtained by fold change and P-value. For each miRNA, a comparison criterion that satisfies P-value <.05 and |fold change| >.05 was selected (P-value: *<.05, **<.01, ***<.001). The 3 dot plots represent fold changes and statistical significances between early and advanced stages (B), alive and dead samples (C), and tumor samples compared to normal (D). Circle size represents statistical significance (P-value was treated by “−log10”).

Figure 4.

Phylogenetic tree of 10 species from UCSC Genome Browser. The number of enriched terms in all 10 species is presented as a heatmap. TE classes for 9 miRNAs are represented in the column annotation bar.

Figure 5.

GO and KEGG terms of 9 miRNA-related genes according to 10 species. In DAVID, enrichment terms were selected for each gene, and GO terms satisfying P-value <.05 were selected.