Comprehensive viral oligonucleotide probe design using conserved protein regions

Jabado, Omar J.; Liu, Yang; Conlan, Sean; Quan, P. Lan; Hegyi, Hédi; Lussier, Yves; Briese, Thomas; Palacios, Gustavo; Lipkin, W. I.

Use of motif finding and set cover minimization markedly increases the computational resources needed to generate probe sets. To determine whether increased complexity results in a more comprehensive probe set, we compared our method to a genome tiling strategy. Probes of 60 nt were designed end-to-end along the entire genome for all 1701 Reference Sequence viral strains available as of May 2007. The tiling probe set served as a contrast to our design method since it was based on nucleic acid sequence, had more probes per gene, required less computation, and included viruses from all genera. In comparison of the methods, the following rules were used to compute database coverage: sequences >400 nt in length were considered covered if six or more probes met hybridization criteria; sequences <400 nt in length were considered covered if two probes met hybridization criteria; sequences <200 nt in length were considered covered with a single probe meeting hybridization criteria. Coverage of the entire database was gauged by computing probe-template ÁG for all 74 044 unique sequence representatives. Database coverage using the tiling method was 47.8% and required 850 136 probes; coverage using the motif-based method was 99.7% and required 270 866 probes (Table 1) .