應用於微流道生物晶片之 多反應物與多目標濃度樣本製備技術

Abstract

樣本製備程序在生物化學應用是一個重要且關鍵的一個步驟。在進行化學反應時，試劑或反應物須以特定的比例混合後方可以進行化學反應。在此程序中， 試劑或反應物經過一連串適當的混合後，來達到反應所需的目標濃度(target concentration)。在樣本製備程序中，如何減少成本一直是大家關注的議題，針對此問題已有數個方法被提出。從先前的研究得知，相較於數位微流體生物晶片僅能支援一比一混合模型，微流道生物晶片能支援更多樣得混合模型，而越多種的混合模型，可以降低製備成本。因此我們想要利用微流道生物晶片的優點來處理多反應物與多目標濃度的樣本製備問題。我們演算法的主要概念是當有多組目標濃度時，讓中介反應物在多組目標濃度中皆能被有效利用，減少廢棄液體，進而達成降低總試劑成本。實驗結果顯示，當同時處理多目標濃度與多反應物樣本製備時，相較於利用既有演算法單獨去處理每個單一目標濃度，我們所新提出來的演算法至多能夠減少26.7%的反應試劑成本。

In most biochemical applications, sample preparation is an essential and crucial procedure. While biochemical reactions take place, reagents or reactants must be mixed in a specific ratio. During this process, different reagents or reactants are mixed in an appropriate sequence to achieve given target concentrations. In the sample preparation problem of how to reduce the cost of reactants is a research topic of concern. To solve this problem, several works have been proposed recently. According to previous works, compared with digital microfluidic biochips using the [1:1] mixing model, flow-based microfluidic biochips with various available mixing models can massively reduce the reactant cost of biochemical applications. Therefore, we are motivated to make good use of flow-based microfluidic biochips to deal with many-reactant and multi-target sample preparation problem. In this thesis, we propose an algorithm that can properly produce intermediate solutions that can be efficiently shared among target concentrations, which effectively reduces the amount of waste solutions and thus minimizes the overall cost during the process. Compared with the prior art that can only process one target concentration at a time, experimental results show that our algorithm can save the reactant cost by up to 26.7%.