TY - JOUR T1 - A data-driven biocomputing pipeline with meta-analysis on high throughput transcriptomics to identify genome-wide miRNA markers associated with type 2 diabetes. AU - De Silva,Kushan, AU - Demmer,Ryan T, AU - Jönsson,Daniel, AU - Mousa,Aya, AU - Forbes,Andrew, AU - Enticott,Joanne, Y1 - 2022/02/02/ PY - 2021/08/27/received PY - 2021/11/23/revised PY - 2022/01/29/accepted PY - 2022/2/16/entrez PY - 2022/2/17/pubmed PY - 2022/2/17/medline KW - Biomarkers KW - Differential expression KW - High throughput transcriptomics KW - Meta-analysis KW - Micro-RNAs KW - Type 2 diabetes SP - e08886 EP - e08886 JF - Heliyon JO - Heliyon VL - 8 IS - 2 N2 - BACKGROUND: MicroRNAs (miRNAs) are sought-after biomarkers of complex, polygenic diseases such as type 2 diabetes (T2D). Data-driven biocomputing provides robust and novel avenues for synthesizing evidence from individual miRNA seq studies. OBJECTIVE: To identify miRNA markers associated with T2D, via a data-driven, biocomputing approach on high throughput transcriptomics. MATERIALS AND METHODS: The pipeline consisted of five sequential steps using miRNA seq data retrieved from the National Center for Biotechnology Information Gene Expression Omnibus platform: systematic review; identification of differentially expressed miRNAs (DE-miRNAs); meta-analysis of DE-miRNAs; network analysis; and downstream analyses. Three normalization algorithms (trimmed mean of M-values; upper quartile; relative log expression) and two meta-analytic algorithms (robust rank aggregation; Fisher's method of p-value combining) were integrated into the pipeline. Network analysis was conducted on miRNet 2.0 while enrichment and over-representation analyses were conducted on miEAA 2.0. RESULTS: A total of 1256 DE-miRNAs (821 downregulated; 435 upregulated) were identified from 5 eligible miRNA seq datasets (3 circulatory; 1 adipose; 1 pancreatic). The meta-signature comprised 9 miRNAs (hsa-miR-15b-5p; hsa-miR-33b-5p; hsa-miR-106b-3p; hsa-miR-106b-5p; hsa-miR-146a-5p; hsa-miR-483-5p; hsa-miR-539-3p; hsa-miR-1260a; hsa-miR-4454), identified via the two meta-analysis approaches. Two hub nodes (hsa-miR-106b-5p; hsa-miR-15b-5p) with above-average degree and betweenness centralities in the miRNA-gene interactions network were identified. Downstream analyses revealed 5 highly conserved- (hsa-miR-33b-5p; hsa-miR-15b-5p; hsa-miR-106b-3p; hsa-miR-106b-5p; hsa-miR-146a-5p) and 7 highly confident- (hsa-miR-33b-5p; hsa-miR-15b-5p; hsa-miR-106b-3p; hsa-miR-106b-5p; hsa-miR-146a-5p; hsa-miR-483-5p; hsa-miR-539-3p) miRNAs. A total of 288 miRNA-disease associations were identified, in which 3 miRNAs (hsa-miR-15b-5p; hsa-miR-106b-3p; hsa-miR-146a-5p) were highly enriched. CONCLUSIONS: A meta-signature of DE-miRNAs associated with T2D was discovered via in-silico analyses and its pathobiological relevance was validated against corroboratory evidence from contemporary studies and downstream analyses. The miRNA meta-signature could be useful for guiding future studies on T2D. There may also be avenues for using the pipeline more broadly for evidence synthesis on other conditions using high throughput transcriptomics. SN - 2405-8440 UR - https://www.unboundmedicine.com/medline/citation/35169647/A_data_driven_biocomputing_pipeline_with_meta_analysis_on_high_throughput_transcriptomics_to_identify_genome_wide_miRNA_markers_associated_with_type_2_diabetes_ DB - PRIME DP - Unbound Medicine ER -