Selection of antisense oligonucleotides on the basis of genomic frequency of the target sequence

Antisense oligonucleotides (ASOs) are capable of blocking the expression of targeted genes and are potential antitumor and antiviral therapeutic agents. The specificity of ASO gene inhibition is compromised when homology to other sequences allows the selected ASO to bind to nontargeted mRNAs. To reduce this nonspecific activity, an ASO should target a sequence that is predicted to be unlikely to occur in other mRNAs. The probability of a sequence being unique can be predicted by determining the genomic frequency of short stretches of sequences contained within the target sequence. Two computer programs, OLIGOMER and HEXAGRAPH, were developed for this analysis. OLIGOMER was used to analyze the genomic frequencies of di-, tri-, and hexamers in more than 24 million nucleotides from 8 different genomes in GenBank. A mathematical model was developed that predicts the genomic frequency of longer oligomers on the basis of the observed frequencies of shorter oligomers. The second program, HEXAGRAPH, was used to graphically display the genomic frequency data of a selected target gene. The computational tools developed in this study may help to design more efficient ASOs by decreasing their nonspecific binding activity.