Quantifying DNA point mutations in commercially available AAV reporter vectors and plasmids

TL;DR

This study compares three DNA sequencing methods to detect mutations in AAV vectors and plasmids, finding inconsistent results across platforms.

Contribution

The study provides a direct comparison of mutation detection accuracy across sequencing methods in AAV products.

Findings

STA-Sanger and Illumina showed similar low mutation rates in vectors and plasmids.

ONT had significantly higher mutation rates but better correlation between vector lots.

Little consistency was found between platforms for mutation location identification.

Abstract

To compare the accuracy of three sequencing methods for quantifying point mutations in AAV products, we sequenced the green fluorescent protein (GFP) transgene from two AAV9 reporter vectors (three lots each) and their respective plasmids, using three sequencing approaches: single-template amplification (STA) followed by Sanger sequencing (STA-Sanger) and two next-generation sequencing (NGS) platforms (Illumina and Oxford Nanopore Technologies [ONT]). Similar vector per-base mutation rates were found for both STA-Sanger (0.016%) and Illumina (0.013%), with slightly lower rates in plasmid sequences (STA-Sanger: 0.0019%; Illumina: 0.0074%), whereas ONT per-base mutation rates were much higher (vector: 2.2%; plasmid: 1.3%). The non-reference majority variant frequency (MVF) was more strongly correlated among vector lots for ONT (R2 = 0.91) compared with Illumina (R2 = 0.46), although correlation between plasmid and vectors was similar for both platforms (R2 = 0.23) and virtually non-existent between platforms for the same sample (R2 < 0.05). Overall, our results showed evidence for single-nucleotide mutations in AAV products, although there was a lack of consistency among sequencing platforms, which underscores the need for the AAV industry to develop sequencing methodologies with improved accuracy as a standard QC protocol.

Rose and colleagues sequenced the transgene of commercially available AAV products using three different sequencing methods. Two methods showed a similar overall per-base mutation rate (consistent across suppliers and production lots) that was higher in the vectors than in the plasmids. However, there was little consistency among methods for identifying the specific location in the gene where the mutations occurred.

Figures