BonoboFlow: viral genome assembly and haplotype reconstruction from nanopore reads

Viral genome sequencing and analysis are crucial for understanding the diversity and evolution of viruses. Traditional Sanger sequencing is limited by low sequence depth and is labor intensive. Next-Generation Sequencing (NGS) methods, such as Illumina, offer improved sequencing depth and throughput but face challenges with accurate reconstruction of viral genomes due to genome fragmentation. Third-generation sequencing platforms, such as PacBio and Oxford Nanopore Technologies (ONT), generate long reads with high throughput. However, PacBio is constrained by substantial resource requirements, while ONT suffers from inherently high error rates. Moreover, standardized pipelines for ONT sequencing encompassing basecalling to genome assembly remain limited.

Here, we introduce BonoboFlow, a standardized Nextflow pipeline designed to streamline ONT-based viral genome assembly/haplotype reconstruction. BonoboFlow integrates key processing steps, including basecalling, read filtering, chimeric read removal, error correction, draft genome assembly/haplotype reconstruction, and genome polishing. The pipeline accepts raw POD5 or basecalled FASTQ files as input, produces FASTA consensus files as output, and uses a reference genome (in FASTA format) for contaminant read filtering. BonoboFlow's containerized implementation via Docker and Singularity ensures seamless deployment across diverse computing environments. While BonoboFlow excels in assembling small and medium viral genomes, it showed challenges when reconstructing large viral genomes.

BonoboFlow and corresponding containerized images are publicly available at https://github.com/nchis09/BonoboFlow and https://hub.docker.com/r/nchis09/bonobo_image. The test dataset is available at SRA repository Accession number: PRJNA1137155, http://www.ncbi.nlm.nih.gov/bioproject/1137155.