Assessing Long-Read Mappers for Viral Genomics

Abstract

Long-read sequencing technologies from Pacific Biosciences and Oxford Nanopore Technologies (ONT) have advanced genomic research, producing reads over 10 kilobases and enabling rapid field-based viral surveillance. This study evaluates eight long-read mapping tools on viral genomic data, including modern and legacy methods. We assessed their performance on ONT reads and their impact on variant calling using bcftools and medaka. Using simulated and real datasets under varying conditions, reflecting different experimental and biological conditions, such as variable read lengths, error rates and the presence of multiple viral vari- ants, we found that the majority of tools had great difficulty in correctly managing read edges. In addition, it was found that with a default setting, the performance of the tools decreased.