Modeling DNA Amplification Technology for Process Optimization

 

Mentor:

Emily Stone, University of Montana.

 

Student Team:

Neil Olver, McGill University

Jon Collis, RPI

Yevgeniy Frenkel, RPI

Derek Moulton, University of Delaware

Ryan Haskett, Duke University

John Evans, RPI

Fritzner Soliman, RPI

Amal Aafif, Drexel University

Dorjsuren Badamdorj, University of Cincinnati.

 

 

Abstract:

 

The objective of this problem is to study PCR reactions in the Ruggedized Advances Pathogen Identification Device (RAPID) Cycler in order to model the reactions and optimize the process of DNA amplification. The RAPID Cycler is a devise designed by Idaho Technologies (IT) to identify DNA sequences using the Polymerase Chain Reaction (PCR) method. This method consists in three basic stages: melting, annealing and extension. In these steps different reactions between the DNA sequences and the reactives take place. These reactions are modeled as sets of ODEs with different reaction rates. The goals in this project are, on one hand, to solve numerically the ODEs in order to fit the reaction rates with the data from IT. On the other hand, the goal is to optimize the time required in each stage of the PCR method, searching for an efficient process. The group modeled the reactions using two approaches. In the temperature-varying model, the reaction rates were treated as functions of the temperature and all reactions occurred simultaneously. In the reduced model, each reaction occurs separatly. An analytic solution of the equations was found for this model. Once the models were formulated, the parameters (reaction rates) were fit using the IT data. To optimize the process, three different approaches were taken. In the fixed time optimization, the idea is to fix the times for each stage in order to minimize the time of the complete process. The stage time optimization allows different times on each stage in order to optimize the efficiency of the process. In the global stage time optimization, each stage is individually optimized. Finally, the Taq-on/Taq-off model was studied.