(451g) Mechanistic Modeling of DNA Hybridization On Surfaces for Improved Microarray Design

DNA microarrays are chip-based, analysis tools which can perform hundreds of thousands of assays in parallel.  They can be used to determine the identity of genes or gene expression levels present in a sample and have been identified as a key technology in genomic sciences and emergent medical techniques.  Despite their abundant use in research laboratories, microarrays have not been used in the clinical setting to their fullest potential due to the difficulty of obtaining reproducible results.  Microarrays work on the principle of DNA hybridization, and can only be as accurate as this process is robust, yet fundamental, molecular-level understanding of hybridization on surfaces is lacking.  Specifically, a detailed knowledge of the various factors affecting hybridization of probes to targets is needed to further refine these devices. 

Using a carefully-parameterized, coarse-grain model of DNA, we have spent the last several years studying the thermodynamics of hybridization on surfaces for simple probe/target pairings.  This has led to several important insights, but a detailed mechanism of the process has never been reported.   In this presentation, we outline efforts to fill in several gaps in understanding.  A detailed description of the hybridization process is presented first.  These results were obtained using extensive computer simulations to create a projection of the free energy of hybridization onto two reaction coordinates: the probe/target separation distance and the angle of approach between the two strands.  Several cases, such as strands of equal and unequal lengths in the bulk and on the surface, are presented.  The free-energy projections were obtained from two-dimensional umbrella sampling, and each required over 11,000 simulations.  Though computationally demanding, this work has created an unprecedented picture of the hybridization process.

In addition to the mechanisms described above, we also present results on how factors found in real microarrays affect the hybridization process.  Specifically, we show how the presence of multiple probes on the surface affects hybridization, and describe how the number and location of mismatched base pairs affect the stability of the duplex.  Finally, we describe how the hybridization process is influenced by the locations of probe/surface connections.  As a whole, the results give the most complete picture to date of the biophysics involved in microarray performance and how this knowledge can be used to improve next-generation devices.