Quantifying the impact of sequencing depth and reference genome choice on metatranscriptomic detection of four bovine RNA viruses.

Abstract

Metatranscriptomic sequencing enables untargeted detection of RNA viruses across clinical and environmental contexts. Evaluating this approach through metrics such as limits of detection and genome coverage, in direct comparison with established molecular tools like qRT-PCR, is essential for understanding its potential diagnostic utility. In the present study, we assessed the detection performance of metatranscriptomic sequencing for four bovine respiratory RNA viruses: Bovine coronavirus (BCoV), Bovine nidovirus (BNV), Influenza D virus (IDV), and Bovine viral diarrhea virus-1 (BVDV-1) in nasal swabs, using reference-based mapping. Sequencing depth (1, 10, and 20 million (M) reads) and reference genome choice (NCBI RefSeq and study-assembled) were systematically varied to quantify their effects on viral read recovery and genome coverage. For BNV and BCoV, sequencing at &#x2265;10&#xa0;M reads was sufficient for metatranscriptomic detection of samples that were qRT-PCR positive at high Ct values (up to 40), but recovering high genome completeness was only achieved for samples with Ct&#xa0;<&#xa0;30. IDV detection was less sensitive: several qRT-PCR-positive samples (Ct 29.6-34.5) yielded no mapped reads even at 20&#xa0;M. BVDV-1 recovery was strongly reference-dependent; mapping to study-assembled genomes markedly increased read counts and coverage compared with NCBI RefSeq, reflecting divergence between field strains and the standard reference sequence. A low number of mapped reads was observed in several qRT-PCR-negative BVDV-1 pools when using the NCBI RefSeq, consistent with potential misclassification of host-derived sequences. By quantitatively linking qRT-PCR Ct values, sequencing depth, and reference divergence, the present study outlines methodological considerations that may guide future applications of metatranscriptomics into veterinary diagnostics.