Algorithms and Software for Oligonucleotide Design
Abstract
Oligonucleotides, or probes, are short, single-stranded fragments of DNA or RNA, designed to bind at a unique position in the target sequence. High specificity of oligonucleotides plays a crucial role in the accurate performance of several important applications such as PCR (polymerase chain reaction) amplification, microarrays, and FISH (fluorescence in situ hybridization) probes. The specificity of existing probes, including commercial ones from Agilent, is lower than expected, usually due to overreliance on the ability of BLAST to find similarities. When probe uniqueness is the goal, the shortcomings of BLAST are most clearly exposed, with bad consequences for the quality of the probes.
We propose two tools for designing probes for whole genome tiling and PCR primers, respectively. The new algorithms rely on the incomparably better ability of multiple spaced seeds to discover similarities, compared with BLAST. They produce more and better probes than existing state-of-the-art tools, the specificity of the new probes being nearly perfect.