In Silico Prediction of Jejunum Permeability by Hierarchical Support Vector Regression
| Autor: | Ming-Han Lee, 李明翰 |
|---|---|
| Rok vydání: | 2018 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 106 Permeability is an important parameter for oral drugs. Intestinal absorption is Closely related to permeability in drugs. Generally speaking, oral drug intestinal permeability measured with two method, in vivo method and in vitro method. Pros of in vivo method are experiment data closer to the reality. Cons are need a lot of time and money. In contrast, Pros of in vitro method are efficient than the first one. Cons are large measurement deviation with unreal situation. An in silico model was produced to quantitatively the oral drug permeability using the hierarchical support vector regression (HSVR) scheme based on the in vivo rat experimental data collected from the literatures(1997-2017). The predictions by HSVR are in good performance with the experiment for those drugs in the training set (n = 53, r2 = 0.93, q2cv = 0.84, RMSE = 0.17, s = 0.07) and test set (n = 13, q2 = 0.75–0.85, RMSE = 0.26, s = 0.14) and outlier (n = 8, q2 = 0.78–0.92, RMSE = 0.19, s = 0.09). Thus, it can be asserted that HSVR can quickly and accurately predict the drug permeability. In addition, the derived model can be adopted as the preliminary metric to carry out biopharmaceutics drug disposition classification system (BDDCS). |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|