Algorithm

The algorithm itself is reasonably straightforward. The decision making process is as follows:

1.

The program walks through the sequence and finds the 40mers with a Tm closest to the target Tm. The twenty 40mers with Tm's closest to the target Tm are kept for further analysis. The program steps through the sequence with a step size proportional to the length of the gene in order to avoid oligos that are very similar packing the top twenty.

2.

Once the twemty oligos with the best Tm are found, other parameters are calculated including: self-complementarity, the heuristics described by Affymetrix, and the Blast score. The oligos are then ranked by Blast score. The motivation behind this is to find probes that will have only one unique hybridization match in the genome.

3.

The program then walks down the list and looks for an oligo which does not have a secondary structure score of 10 or more and passes the heuristic described by Affymetrix. If none of the oligos pass both of these tests, the oligo with the lowest Blast score is chosen by default.

Last updated on 7 December 1999 Maintained by sugnet@cse.ucsc.edu