本文標記轉錄因子, 重複序列和工具預測出的黏合序列定位於基因前的促進區域。應用資料探 (Data Ming) 技術於重複序列與轉錄因子的組合以及工具預測出的黏合序列與轉錄因子的組合。再從關聯規則中去除多餘的規則．利用統計方法找出較有意義的，在規則裡的重複序列和工具預測出的黏合序列中找尋可能的轉錄因子。由於不同的轉錄因子組合的黏合會造成基因的轉錄有所不同，因此我們找出不同功能之相關基因較具鑑別性的組合。我們進行的實驗主要是酵母菌及原蟲的基因組上。轉錄因子的研究上，我們得到相當有價值的資訊，並將結果公開在http://dbms68.csie.ncu.edu.tw/REDB/ 網站上。 The data mining approach, mining association rules, is applied to mine the associations from the combinations of candidate regulatory sites and known regulatory sites. We apply a set of statistical algorithms to characterization of the site combinations in a co-regulated gene group and statistically analyzed it to other co-regulated gene groups to find the site combinations which prefer to occur in a specific gene groups with significant occurrences. The regulatory sites of the gene group-specific site combinations are putative transcription factor binding sites. The methodology introduced here facilitates to analyze combinatorial interactions of multiple transcription factors and is applied to two organisms, Saccharomyces cerevisize and Caenorhabditis elegans, and the promoter regions of ORFs of them. The results are now available at http://dbms68.csie.ncu.edu.tw/REDB/
