Structural analysis of recombinant AAV vector genomes at single-molecule resolution
David Rouleau, Dimpal Lata, Serena Dollive, Robert E. Bruccoleri, Laura Van Lieshout, Diane Golebiowski, Ifeyinwa Iwuchukwu
Abstract
Recombinant adeno-associated virus vectors are essential tools for in vivo gene therapy, yet heterogeneity in their packaged genomes remains an important safety consideration. To systematically evaluate this heterogeneity, we developed a long-read, read-level analysis pipeline that directly classifies individual AAV genomes and their structural variants from PacBio sequencing data.
Introduction
Recombinant adeno-associated vectors (rAAVs) are emerging as a valuable tool for their use in gene therapy, with several candidates showing promising clinical trial outcomes and some already approved by the U.S. Food and Drug Administration [1–3]. At the same time, clinical experience has highlighted important safety concerns including hepatotoxicity, thrombotic microangiopathy, myocarditis, and hepatocellular carcinoma [4–7].
Materials and method
2.1. AAV vector sample production
The vectors analyzed in this study were recombinant GFP constructs, including a self-complementary GFP (scGFP) containing a genome of approximately 2080 nucleotides (nt) with a payload size of 1831 nt, and a single-stranded GFP (ssGFP) containing a genome of approximately 2162 nt with a payload size of 1872 nt. Both ssAAV and scAAV genomes expressed GFP using the CB promoter (CMV enhancer, chicken beta actin promoter, SV40 intron) and bGH polyA signal.
Results
3.1. In silico testing
To validate the accuracy of our structural variant calling pipeline, we used our own sequence generator program to produce 68 files of 10,000 AAV genome structural variants sequences and ran them through our structural variant calling pipeline. The files consisted of four replicates of 17 structural variant sequence files, each with a different level of error applied to its sequences.
Discussion
Adeno-associated viral vector (AAV) genome heterogeneity remains a critical concern in gene therapy vector design and quality control. While previous studies have explored the diversity of packaged genomes through various methods, including short-read NGS, they have been unable to reveal the complete structure of entire packaged vector genomes within an AAV sample [47].
Conclusion
Developing long-read sequencing technologies has given researchers the ability to fully sequence vector genomes. Although numerous structural variants have been reported in the literature, there remains a limited number of bioinformatics workflows specifically designed to comprehensively analyze rAAV genome heterogeneity.
Acknowledgments
We thank James McGivney for his support in establishing AAV sequencing capabilities. We are also grateful to Kiran Adhikari and Celia Slater for their assistance in reviewing the manuscript, and to Brenda Burnham, Evan DaSilva, Ben Rogers, Matt Perez, Brian Brazell, and Bryan Alighieri for performing the vector QC analyses. We further acknowledge the team at Pacific Biosciences for their expert technical assistance and for ensuring reliable performance of the Sequel II sequencing platform.
Citation: Rouleau D, Lata D, Dollive S, Bruccoleri RE, Van Lieshout L, Golebiowski D, et al. (2026) Structural analysis of recombinant AAV vector genomes at single-molecule resolution. PLoS One 21(7): e0339201. https://doi.org/10.1371/journal.pone.0339201
Editor: Vahid Mansouri, Tehran University of Medical Sciences, IRAN, ISLAMIC REPUBLIC OF
Received: December 2, 2025; Accepted: July 10, 2026; Published: July 30, 2026
Copyright: © 2026 Rouleau et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The parsing codebase, the data analyzed, and the sequence generation code used to generate in silico sequence data is available on Zenodo at https://doi.org/10.5281/zenodo.19862292 to support reproducibility and community use. The code and data used for breakpoint analysis is also available on Zenodo at https://doi.org/10.5281/zenodo.20072387.
Funding: The author(s) received no specific funding for this work.
Competing interests: I have read the journal’s policy and the authors of this manuscript have the following competing interests: All authors were employed at Oxford Biomedica (US) LLC and predecessor companies during their contributions to the manuscript.
