Combinatorial Design of Universal DNA Tag Systems
 

CPSC 536A Course Project
Spring 2001
 
 
Tim Braun tbraun@cs.ubc.ca
James Gauthier gauthier@cs.ubc.ca
Tam Huynh tam@juggler.net
Louie van de Lagemaat lvandela@terryfox.ubc.ca



 

Introduction
 

An important task in biotechnology is to analyze a mixture of DNA strands in order to investigate certain properties of their structure and composition.  In most cases, such a mixture may contain thousands of types of DNA strands.  Therefore, DNA strand analysis requires fast, yet reliable, methods to probe the strands.  One technique is known as array-based hybridization assay.  Conventional array-based hybridization assays involve probing DNA in solution with probes attached to a chip in a regular rectangular array.

To address the need for a multitude of different arrays, Ben-Dor, et. al. discuss a universal Tag/AntiTag (TAT) system.  In this system, the probe for a given gene or DNA strand of interest is ligated to a tag, the antitag for which has been synthesized on the array.  This tag/probe construct is used as a reporter molecule.  Thus, in practice, an experiment consists of two steps.  In the first (solution-phase) step, mixed single-stranded DNAs of interest hybridize to reporter molecules.  In the second step, the reporter/sample pairs are washed over the array and attach to their respective antitags.  Fluorescent markers attached to the DNAs of interest give a measurable signal.

Theoretically, DNA tags should match only to their antitags.  Thus, to avoid spurious cross-hybridization, tag/antitag pairs must be distinct.  This problem presents an interesting challenge.  Ben-Dor, et. al. present a combinatorial solution to this challenge.  This solution involves a simplistic computation of a melting pseudo-temperature for each tag/antitag pair based on base-pairing. The melting temperature is calculated as two points for each CG base pair and one point for each AT pair (this is known as the 2-4 rule).  Combinations of tags are limited by eliminating cross-hybridizations of melting temperature higher than C, a ``cool'' temperature, and only allowing true hybridizations of melting temperature higher than H, a ``hotter'' temperature.

Our goal with this project is to implement an algorithm similar to the one presented by Ben-Dor, et. al.  However, we would like to use a more sophisticated estimate of melting temperature than the simplistic 2-4 rule.  Research has led us to consider the Nearest Neighbour Model as an estimate of melting temperature.
 
 

Contents
 


 

Updated April 4, 2001