應用於微流道生物晶片上多環節混合器以達到反應物最小化之樣本製備技術

Abstract

樣本製備程序(sample preparation)為各種生化反應中不可或缺的步驟之一。原始生物樣本或反應試劑必須在此程序中進行稀釋或混合，以達到反應所需的目標濃度 (target concentration)。現有的樣本製備程序演算法大多僅使用(1:1)之混合模型進行兩溶液混合。因此，本篇論文提出全球第一篇使用多種混合模型以及多溶液混合之樣本製備程序演算法。此演算法藉由混和器內反應物含量分析技術 (volume-oriented)達到所需目標濃度，簡稱VOSPA演算法。在單一目標樣本製備流程中，VOSPA藉由分析溶液內樣本含量，並在混合過程中有效利用中間產物，產出能節省試劑使用量以及減少反應次數的樣本製備程序。在多目標樣本製備程序中，VOSPA-M 藉由預知尚未產生之廢液，利用有向無環圖確保混合步驟的相依性，提前將未來廢液規劃再利用以有效減少反應物使用量。實驗結果顯示，VOSPA 及 VOSPA-M 可有效降低反應物用量，達到更佳的樣本反應製備流程。

Sample preparation is one of essential processing steps in most biochemical reactions. In this process, raw reactants are diluted to achieve given target concentrations. So far, most of existing sample preparation techniques only consider mixing of two source solutions under the (1:1) mixing model. In this thesis, we propose a volume-oriented sample preparation algorithm, VOSPA, for sample preparation on flow-based microfluidic biochips. VOSPA is the first sample preparation algorithm that not only blends multiple (≥2) concentration values in a mixing operation but also allows various mixing models. For single-target sample preparation, VOSPA enables segment-based intermediate solutions reuse for better reactant minimization by tracking wasted solutions in the CV Bank. For multi-target sample preparation, VOSPA-M adopts a look-ahead bank strategy to reuse those foreseeable wasted solutions associated with target concentrations. Experimental results show that VOSPA and VOSPA-M are better solutions for reactant minimization as compared with the previous works.