| Summary: | 碩士 === 國立陽明大學 === 生物醫學資訊研究所 === 105 === When investigating cancer markers, it is likely to get a lot of potential candidates, and the number of candidates is very large. Researchers may be based on their own experience and know-how to select the more likely marker candidates for further study. This process is costly and time consuming, so we hope to find a way to prioritize marker candidates. In the determining cancer markers, the expression level on tissues is one of the main considerations. Previously, we developed a prioritization method, HPA Scoring, based on tissue expression annotations deposited in the Human Protein Atlas (HPA). In this study, we updated our dataset to the latest version of HPA and adjusted the scoring method to the following: 1) using median instead of average in calculating caner expression (EiC); 2) normalizing expression differences (ED) according to ranks. As a result, we calculated the scores for all 20,783 antibodies involving 15,297 genes, and obtained 548,807 scores covering 20 different types of cancer. The adjusted scores will be compared to the previous calculation. The number of markers will be verified according to cancer marker annotations retrieved from another independent dataset, Ingenuity Knowledge Base. We hope that our results can be achieved: First, the new score and the original version of the comparison, the top 100 in the IPA to determine the numbers of markers are more than the original version. The results show that the number of not less than the original ratio of 78%. Second, the expression in cancer tissue is high. Third, the expression is large difference in the cancer tissue and normal tissue.
|
|---|
