Complete viral RNA genome sequencing of ultra-low copy samples by sequence-independent amplification

Malboeuf, Christine M.; Yang, Xiao; Charlebois, Patrick; Qu, James; Berlin, Aaron M.; Casali, Monica; Pesko, Kendra N.; Boutwell, Christian L.; DeVincenzo, John P.; Ebel, Gregory D.; Allen, Todd M.; Zody, Michael C.; Henn, Matthew R.; Levin, Joshua Z.

As the Illumina reads generated from RNA processed with the Ovation RNA-Seq kit start with a variable number of bases derived from the kit primers, the first 18 bases of each read were trimmed prior to alignment. To analyse input data composition, we first employed an efficient Illumina aligner BWA-v0.5.9 (26) by aligning reads to viral and host reference genomes. For HIV clone and clinical samples, the reference genomes used were the concatenation of the human genome assembly (NCBI36/hg18), human rRNA sequences (NR_003286.1, NR_003287.1, V00589.1, NR_003285.2, gij251831106: 648-1601, gij251831106:1671-3229) and the HIV HXB2 sequence (K03455.1). For WNV clone samples, the reference genomes used were the concatenation of preliminary and unpublished sequencing data from the golden hamster, Mesocricetus auratus, (K. Lindblad-Toh, personal communication), hamster rRNA sequences (NR_045212.1, NR_045133.1, NR_045213.1, D89009.1, DQ334843.1) and WNV NY99 reference (NC_009942.1). For RSV clinical samples, the reference genomes used were the concatenation of the human genome assembly (NCBI36/hg18), human rRNA sequences (NR_003286.1, NR_003287.1, V00589.1, NR_003285.2, gij251831106: 648-1601, gij251831106:1671-3229) and the RSV A sequence (M74568.1). The alignment was carried out by first aligning each read independently using command: bwa aln (-q 5 -l 32 -k 2 -t 4 -o 1), then the read pair information was used by invoking command: bwa sampe (-a 100 000). MergeBamAlignments, from the picard package (v1.59) [http://picard.sourceforge.net/], were used to return the unmapped reads to the aligned BAM file. A custom script was used to count the number of reads aligning to each of the concatenated references. Compared with the BWA aligner, the Mosaik aligner (version 1.1.0013) (http://bioinformatics.bc.edu/marthlab /Mosaik) is slower but more accurate when aligning relatively long Illumina reads like the ones used in the current studies. To determine more accurate coverage information for our viral samples only, we used Mosaik (MosaikAligner -st illumina -hs 10 -act 15 -bw 29 -mmp 0.25 -minp 0.25) to align reads to the target viral reference genome and the consensus assembly results. Primer sequences were not trimmed since soft-clipping is tolerated by Mosaik.