RNA virus discovery.

Sequencing reads were demultiplexed and trimmed for quality with Trimmomatic (40) before de novo assembly using Trinity (41). The resulting contigs were first compared against the database of all reference RNA virus proteins downloaded from GenBank by using BLASTX with an E value cutoff at 1E−5. Potential viral contigs were then compared to the entire nonredundant nucleotide (nt) and protein (nr) database to remove false-positive results. The quality-filtered virus contigs with unassembled overlaps were then merged by using the SeqMan program implemented in the Lasergene software package v7.1 (DNAStar). To confirm the assembly results, reads were mapped back to the virus genomes with Bowtie2 (42) and inspected by using Integrated Genomics Viewer (IGV) (43) for any assembly errors. The final sequences of the virus genomes were obtained from the majority consensus of the mapping assembly.