Ball Python Nidovirus: a Candidate Etiologic Agent for Severe Respiratory Disease in Python regius

Stenglein, Mark D.; Jacobson, Elliott R.; Wozniak, Edward J.; Wellehan, James F. X.; Kincaid, Anne; Gordon, Marcus; Porter, Brian F.; Baumgartner, Wes; Stahl, Scott; Kelley, Karen; Towner, Jonathan S.; DeRisi, Joseph L.

We then searched for possible pathogen-derived sequences in what remained of the data sets. We performed BLASTx alignments using the remaining reads to search a database of virus protein sequences. In case samples but not in control samples, we observed sequences with similarity to viruses in the Nidovirales order. The data set from snake 2 had the most nidovirus-like sequences: 92 read pairs with BLASTx alignments to nidovirus protein sequences with an expect (E) value of Յ10 Ϫ3 (see Table S2 in the supplemental material). These sequences were used to seed targeted de novo genome assembly using the PRICE metagenomic assembler (22) , which produced a draft assembly of what appeared to be the complete viral genome of 34.5 kb (Fig. 4A ). Sanger sequencing of the entire genome and rapid amplification of cDNA ends (RACE) were used to validate the assembly and to confirm that it contained the proper end sequences (Fig. 4B ) (23) (24) (25) . We used the Bowtie2 software alignment tool to retrospectively map paired-end reads to the draft genome assembly and found it to be well supported by deep sequencing data (26, 27) . Coverage levels were even across the genome, and read pairs generally mapped concordantly ( Fig. 4C and D) . We found that the deep-sequencing-based assembly represented nearly the entire genome, except for 165 bases of 3= untranslated region (UTR) sequence, which were determined by 3= RACE. This genome sequence has been deposited in GenBank (see below).