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.

The most recent Pfam-A release (Version 22) comprised 9318 families, of which 1540 had viral members. Of 405 543 annotated protein sequences with length >20 aa, 278 119 (68.6%) belonged to a Pfam-A family, while 127 424 (31.4%) did not. Three probes were chosen for each gene, yielding a total of 104 467 cPf and 133 513 cNPf probes. Of sequences not contained in Pfam-A, only 5.6% (6956) were found in Pfam-B alignments. Thus, due to the lower quality of alignments (23) and poor viral representation, the Pfam-B was not used for probe design. The 12 428 untranslated regions processed yielded 4616 probes. For the 24 841 unannotated sequences processed, 13 740 probes were designed. Sequences that were not covered due to high/low GC%, low complexity, repetitive sequence or a preponderance of ambiguous nucleotides (4244) were processed with a sliding window strategy; 14 530 probes were designed. Overall, the number of probes required to address all viral sequences was 270 866. Sequence counts and probe counts for the most recent EMBL/Pfam release are detailed in Figure 4 . An example of typical probe distribution is shown with respect to the Dengue virus 1 genome (NC_001477; Figure 5 ).