| 申請系所(單位) |
獸醫學院獸醫系 |
| 計畫主持人 |
張佳瑜助理教授 |
| 計畫名稱(中文) |
使用機器學習設計親和度最佳化之重組抗體及應用其開發疾病診斷試劑 |
| 計畫名稱(英文) |
Machine Learning-Driven Affinity-Optimized Recombinant Antibody Design for Disease Diagnostic Assays |
| 共同主持人 |
1. 機械工程學系(所) 藍國瑞助理教授 2. 生醫工程研究所 賴千蕙副教授 |
| 協同主持人 |
|
| 中文摘要 |
重組抗體在疾病診斷、醫療措施、基礎研究方面皆為不可或缺的生物材料。在傳染病領域，抗體可作為治療工具以控制感染，亦可應用於診斷試劑之開發，以快速檢測特定病原。然而傳統抗體製備與篩選的方式研發週期長、耗時耗力。且傳統方法從成千的抗體產製細胞中篩選出具有特定抗原辨識的抗體較為隨機且無效率，抗體與抗原之間的親和力與特異度往往難在第一時間進行準確評估與最佳化，因而導致後續相關應用開發的拖宕。為加速抗體篩選與製備，本計畫將結合人工智慧機器學習輔助生產親和力最佳化之重組抗體，建立一個可縮短抗體製程與增加篩選信心的平台，並進行試驗驗證與比對最佳化的抗體對抗原的親和力與特異度。除此之外，本研究亦欲將該抗體嘗試不同化學修飾，應用於發展微流體晶片與快篩診斷試劑等，增進抗體於疾病診斷平台的泛用性。 |
| 英文摘要 |
Recombinant antibodies are indispensable biological materials in disease diagnosis, medical applications, and basic research. As to infectious diseases, antibodies can be used in therapy to control infections and applied to diagnostic tools for efficient disease diagnosis. However, traditional approaches for antibody preparation and selection are time-consuming and labor-intensive. Moreover, the screening for target antibodies from thousands of candidates is rather random and the binding affinity and specificity of antibody are often difficult to evaluate and optimize in time. Taking advantage of the advanced development of artificial intelligence, machine learning models will be introduced herein to facilitate the production and selection of the recombinant antibodies with optimized affinity. Subsequently, the recombinant antibodies will be chemically modified for further applications such as microfluidic chips and rapid tests. |
