建置微小核醣核酸生物標記資料庫

Abstract

近年已發現微小核醣核酸(miRNA)與許多種癌症與疾病有關。miRNA 是長度大約為22個核苷酸的非編碼微小RNA分子，其功能為在轉錄後修飾時藉由抑制基因表現而導致許多癌症和疾病。許多研究確認miRNA在不同組織與不同型態細胞中，其表現差異將造成生物體健康危害與特定疾病的形成，其中包含各種慢性疾病，心血管疾病與各種癌症，例如miR-205已被許多研究證實與早期卵巢癌有顯著的相關性。因此大規模蒐集、與建構miRNA與已知疾病及癌症成因之生物性資料庫，並藉此分析miRNA與疾病及癌症之相關性，將能藉此找出可能應用於臨床診斷與治療疾病及癌症的生物標記。由於眾多的研究資料探討miRNA與疾病之間的關係，為了更有效的整合與分析以找出與疾病及癌症相關之生物標記，本研究開發了miRBiomarker資料庫(網址 http://140.113.239.240/~yubau)。此資料庫利用text-mining技術收集2016年五月之前3萬篇miRNA及疾病相關之研究報告最後篩選審核輸入了，包含687篇文獻，166種不同文獻所描述的疾病，789筆miRNA 生物標記資訊，並整合TCGA 資料庫miRNA表現量並呈現視覺化表現量圖表，以人工閱讀方式記錄所有在研究生物標記與疾病及癌症相關性時所需要的重要資訊，以利未來的整合統計分析所需。本研究所建立之miRBiomarker資料庫，預期將在尋找臨床診斷與探索疾病與癌症成因相關之生物標記時有很大的幫助。

miRBiomarker (available via the Web site at http://140.113.239.240/~yubau), a manually curated database, offering a comprehensive resource of biomarker among various human diseases. MicroRNAs (miRNAs) are 18-24 nucleotide RNA, which play a very important role in the part of post-transcriptional regulation of genes in animals and plants. They regulate gene expression by repressing mRNA translation or by degradation of target sequence. MicroRNAs are also useful in diagnosis of diseases. Previous studies have reported that miR-205 is strongly related with early stage ovarian cancer. However, current miRNA databases do not contain enough human disease biomarker information. It is necessary to build the most comprehensive biomarker database for miRNAs and human disease with a user-friendly web interface. Thus, we construct a miRBiomarker database that includes miRNAs expressions from The Cancer Genome Atlas (TCGA). Also, it collects thousands of miRNA entities by manually reading related literature that is identified using text mining method. The current version of miRBiomarker documentsed 789 verified biomarker entities in 166 different human diseases by reviewing 687publication. All the important information of biomarker including miRNA ID, disease name, expression pattern, a brief description of the biomarker, sample type, diagnostic and prognostic value are provided in details in the database. We would like to provide more user friendly and effective web interface functions for users to find candidates of miRNAs and their related diseases. Therefore, miRBiomarker could support researchers to identify miRNA biomarkers and become one of the major tools in this field.