A Digital Microfluidic Biochip (DMFB) offers a promising platform for medical diagnostics, DNA sequencing, Polymerase Chain Reaction (PCR), and drug discovery and development. Conventional Drug discovery procedures require timely and costly manned experiments with a high degree of human errors with no guarantee of success. On the other hand, DMFB can be a great solution for miniaturization, integration, automation, and cost reduction of drug discovery. DMFB can improve the parallelism of drug discovery procedures; since most procedures in drug discovery are concurrent and parallel, DMB can reduce the execution time of these bioassays. Therefore, there is a critical need to develop DMFBs to speed up the drug discovery applications and improve cost and error of these reactions. In this paper, a new architecture is used for drug discovery applications. The architecture is evaluated and compared with FPPC architecture. The experimental results prove that the new architecture is faster and cheaper than FPPC; it reduces all the important parameters such as total execution time, number of controlling pins, CAD algorithm execution time, and the area usage and its costs. There is an urgent need for collaboration between experts of drug discovery, microfluidic platform architecture and also machine learning to design a data-driven microfluidic architecture which improves the CAD algorithms by learning from prior knowledge.