Make it so: Rapid and affordable plasmid sequencing on ONT platforms with PICARD-seq

Plasmid construction is a fundamental aspect of molecular and synthetic biology, yet current validation methods often focus solely on the inserted fragment rather than the entire plasmid structure. Approximately one-third of laboratory-made plasmids contain sequence errors that can impact their functionality. Traditional Sanger sequencing becomes impractical for validating large plasmids due to its slow speed, while short-read sequencing methods struggle to accurately resolve repetitive DNA regions, such as promoters, that are prevalent in synthetic constructs.

In this study, researchers introduce PICARD-seq, a rapid nanopore sequencing protocol that employs readily available Tn5 rapid barcoding reagents and a MinION device. This method allows for the sequencing of entire plasmid pools within a single day. The study thoroughly evaluated the computational tools used to analyze the resulting nanopore sequencing data.

The researchers tested five independent replicates of each pipeline using a curated set of 25 plasmids with known sequences and lengths ranging from 3.0 to 20.6 kilobases. These plasmids included various dilution series and repetitive multi-part constructs to simulate real-world scenarios.

Among the tested nanopore workflows, the ONT EPI2ME Clone Validation method proved to be the fastest, completing its task in 13-18 minutes, but it exhibited high variability across replicates, both in terms of assembled plasmid sequences and the specific sequences returned. In contrast, the Canu assembler outperformed the default Flye assembler, demonstrating superior performance in assembling large, repetitive, and dilute plasmid samples.

Additionally, reducing the minimum coverage parameter from 60 times to 20 times significantly enhanced the assembly quality for these challenging samples. However, the ensemble assembler Autocycler, although slower, ranging from 81 to 111 minutes, yielded the most consistent results and the highest rate of recovering the expected plasmid sequences.

Further analysis using minimap2 for read mapping helped distinguish true sequence differences from assembly artefacts. When applied to problematic plasmids, PICARD-seq was able to detect common issues such as backbone concatemers, misincorporated promoter regions, and mixed populations of rearranged molecules within repetitive constructs.

Overall, PICARD-seq makes the routine, affordable, and practical validation of whole plasmid structures accessible to individual laboratories, overcoming the limitations posed by traditional sequencing methods.