Defective But Promising: Evaluating Bioinformatic Pipelines for Utility of Defective Interfering RNA Discovery in Plant Viral Infections

We explored the utility of the currently available bioinformatics programs ViReMa, DI-tector, DVGfinder, DG-Seq, and VODKA2 for identifying junction points in plant virus high-throughput sequencing (HTS) data that could be tested downstream for antiviral capacity. Specifically, we looked at whether the outputs from these bioinformatic tools generally agree and whether the most frequently identified "defective viral genomes" (DVGs) from these programs are promising defective interfering RNA (DI) candidates for downstream validation. We also explored the possibility of these tools helping us address a larger research question of whether DI RNA are consistently generated and maintained in a specific virus-host combination when conditions are permissive for their replication and accumulation, our "DI prevalence" hypothesis. This was conducted by running eight previously published RNAseq datasets through all five programs and comparing degree of output overlap, most common junction point identified, and whether previously published DI junction points were found. Our results demonstrate a low degree of agreement regarding identified junction points between programs, promise regarding looking at the most commonly occurring junction for DI candidates, and support for our DI prevalence hypothesis. We conclude that bioinformatics workflows have a place in the toolbox of DI and DVG research, but they should not be used alone. We suggest the use of multiple programs on a dataset to better inform decisions regarding deletions to re-create and screen downstream and reiterate the importance of other avenues of evidence in DVG/DI characterization.