OligoArray 2.0: design of oligonucleotide probes for DNA microarrays using a thermodynamic approach

Abstract

There is a substantial interest in implementing bioinformatics technologies that allow the design of oligonucleotides to support the development of microarrays made from short synthetic DNA fragments spotted or in situ synthesized on slides. Ideally, such oligonucleotides should be totally specific to their respective targets to avoid any cross-hybridization and should not form stable secondary structures that may interfere with the labeled probes during hybridization. We have developed OligoArray 2.0, a program that designs specific oligonucleotides at the genomic scale. It uses a thermodynamic approach to predict secondary structures and to calculate the specificity of targets on chips for a unique probe in a mixture of labeled probes. Furthermore, OligoArray 2.0 can adjust the oligonucleotide length, according to user input, to fit a narrow T(m) range compatible with hybridization requirements. Combined with on chip oligonucleotide synthesis, this program makes it feasible to perform expression analysis on a genomic scale for any organism for which the genome sequence is known. This is without relying on cDNA or oligonucleotide libraries. OligoArray 2.0 was used to design 75 764 oligonucleotides representing 26 140 transcripts from Arabidopsis thaliana. Among this set, we provide at least one specific oligonucleotide for 93% of these transcripts.

Figure 1

Distribution of melting temperature of 15mer oligonucleotides with various GC contents. For each GC content (from 5 to 10 G or C over 15 nt), 1000 random sequences have been generated and their T_m calculated and sorted. (A) For each GC content, the T_m distribution has been plotted. The two vertical lines visualize the position of the 26th and 974th T_m values and define the 2.5% of data excluded at each extremity of the curves (total 5%). (B) The T_m values from the same data set have been plotted as a function of the GC content of the oligonucleotide (number of GC over 15 nt). The gray squares represent the 26th and 974th T_m values from the distribution.

Figure 2

Minimum and maximum melting temperatures as a function of the length and GC content of oligonucleotides. For each oligonucleotide length between 15 and 70, with an increment of 5 nt, and for each GC content between 20 and 80% of GC, the minimum and maximum valid T_m have been plotted. The upper plane defines the maximum T_m that can be computed for a given length and GC content. The lower plane defines the minimum T_m.