Twenty Years of Sharing: Addgene's Viral Vector Service

As part of Addgene’s 20^th anniversary celebration, we’ve been finding and sharing stories unique to our repository and resources. As editor of Addgene’s eBooks — including the recently released Viral Vectors 101 2^nd edition — I was incredibly excited when we picked our viral vector service as one of those stories. Viral vectors are viruses engineered to be safely used to deliver a genetic payload to cells, and the complexity of effectively deploying them in the lab somewhat mirrors the complexity of understanding their wild counterparts.

Those familiar with viral vectors, and especially those familiar with adeno-associated viruses (AAVs), will know just how tricky these tools can be to use. AAVs, which can target different cell types depending on their serotype, are particularly challenging. Their efficacy in any given system can be affected by any number of factors, including the age of the model organism, injection site, and titer.

Addgene’s Viral Vector Service

Besides being complex to use, viral vectors are also difficult both to produce and to verify the quality of the prep. This was a large factor in Addgene’s decision to launch our viral vector service in 2016. We realized we could apply the lessons learned from our plasmid lab to develop a reliable and robust high quality viral vector service, making an array of ready-to-use viral vectors available to researchers around the world. While we still offer many resources for producing vectors from plasmids — including the protocols we use in our lab — being able to order high quality viral vectors from the Addgene repository can save researchers significant time and removes a variable from the lengthy process of validating and using an AAV in a specific experimental setup.

Teams across Addgene have worked hard to ensure that the service has grown steadily, helped by a partnership with the University of Pennsylvania’s Vector Core that began in 2018. By the end of 2023, our catalog had grown to 879 viral catalog items (Figure 1). The service caught on quickly in the scientific community, turning slightly under ten requests a day in 2017 to over fifty requests a day in 2023. And while viral vectors are not as easy to ship as plasmids, they still have a global reach. From 2016–2023, we shared 93,491 viral vector preps with researchers in 45 different countries — and as of 2024, we’ve distributed over 100,000 viral vector preps!

Figure 1: A) Cumulative AAV preps available in the Addgene repository each year from 2016–2023. B) AAV preps distributed per year by Addgene from 2016–2023.

Tools and targeting

The amount of vectors available and number of requests tell only part of the story. Viral vectors are, after all, a delivery method. I would be remiss to talk about them without looking at what they’re delivering and where they’re delivering their payload to.

Figure 2: Numbers of AAV tools, by category, available from Addgene as of 2023.

A large percentage of viral vector preps contain biosensors, optogenetic tools, and chemogenetic tools: protein tools that are activated, respectively, by substrates, light, or engineered ligands (Figure 2). Biosensors, which include calcium and voltage sensors, are the broadest category of viral vector tools, which may help explain why they are the most commonly requested (Figure 3). Over 400 biosensors were distributed in 2023 alone. You’ll also find a large number of requests for vectors in the “controls” category, which includes things like viral vectors that deliver GFP.

Figure 3: A) Cumulative AAV preps available at Addgene from 2016–2023 by type of tool. B) AAV preps distributed by Addgene per year from 2016–2023 by type of tool.

Equally as important to what is being delivered is where it’s being delivered. Since target specificity is determined by the capsid of the AAV, many of the vectors in Addgene’s catalog are offered in different serotypes. In fact, the 879 viral vectors in our repository are produced from only 413 plasmids containing a genetic payload. While AAV9 and AAV1 are the most requested serotypes (Figure 4), our catalog includes vectors in a variety of serotypes, including the retrograde AAVrg and the systemic capsid PHP.eB.

Figure 4: A) Cumulative AAV preps for the seven most popular serotypes available from Addgene from 2016–2023. B) AAV preps for the seven most popular serotypes distributed per year by Addgene from 2016–2023.

How to grow a catalog

If you’ve been following our blog, you may recall our first 20^th anniversary story on CRISPR, where we dove fairly deeply into the different types of tools scientists were using/depositing. Our viral vectors data is less amenable to such analysis, due to its inherent bias: Addgene uses a prioritization process to decide which tools we make available.

This prioritization happens for good reasons. We began offering our viral vector service because we realized that having reliable ready-to-use vectors made from plasmids in the Addgene catalog would be broadly useful to the research community. It does, however, necessitate us first identifying broadly useful vectors, a process that isn’t as simple as it seems.

Take, for instance, the GCaMP-encoding viral vectors available in the repository. GCaMP, a calcium indicator (biosensor, by Addgene’s categorization) was first developed in 2001, with successive generations developed ever since. GCaMP6 was offered through our viral vector service in 2016, and it quickly became one of our most popular ready-to-use viral vectors, making up almost 10% of our requests by 2019 (Figure 5A). We then added jGCaMP7 and eventually jGCaMP8, later generations of GCaMP6, to the service. While they were both requested, GCaMP6 remains the more requested option, and jGCaMP8 appears to be growing in requests faster than jGCaMP7 (Figure 5B).

Figure 5: A) GCaMP6, jGCaMP7, and jGCaMP8 viral preps distributed per year compared to other AAVs preps from 2016–2023. B) Inset showing distribution of only GCaMP6, jGCaMP7, and jGCaMP8 per year from 2016–2023.

All three generations are regularly requested and a valuable part of our viral vector service. But why are their request patterns so different? It could be a matter of timing, as jGCaMP7 was released relatively soon after GCaMP6. It could be a matter of the size of the incremental improvements between generations: jGCaMP8 allows for the “picture” of a train of action potentials that GCaMP6 and jGCaMP7 do not. Most likely, it’s a combination of many factors unique to this set of tools and the community they serve.

As difficult as it may be to find answers to such questions, Addgene is deeply interested in pursuing them. We’ve recently revamped our viral vector selection process to standardize our feedback and evaluation of viral vectors. This process builds on our former approach and involves reading primary literature, soliciting information from requestors and depositors, and identifying trends in our own repository data, giving us significant insight into which vectors are likely to be of most use to the community. Thoughtfully prioritizing which preps to include in our service allows us to make ready-to-use viral preps that are available quickly.

Beyond the service

This brings us to the final piece of the viral vectors service puzzle: the usefulness of each viral prep in practice. The number of requests shows how many scientists believe a vector would be useful in their experiment…but, even for our blue flame viral preps, it doesn’t actually show you if the vector was useful for the requestors or how it performed.

Addgene launched our AAV Data Hub as a community platform for the sharing of that exact data. Here, requestors can upload their experimental data in a standardized format, sharing how well a given viral vector worked — or didn’t work — in their system. The data is curated and published with a DOI, allowing others to reference and cite the data when deciding on which viral vectors to validate in their own systems. Browsing the Data Hub before selecting an AAV can help a researcher make a more informed choice on which AAV to request and/or save time in selecting experimental conditions.

Sharing our experiences

Addgene’s plasmid repository gave us the experience to launch a high-quality ready-to-use viral vector service. But in the years since, we’ve learned so much about what it takes to make these viral vectors not only available, but accessible. The supporting data and community engagement required to do this drove Addgene to develop new systems, new partnerships, and new platforms, from our early engagement with UPenn to our launch of the redesigned Data Hub in 2022. It’s influenced the way we approach our viral vector educational resources, from eBooks to guides to protocols, which was why I was so excited to delve into this data as we updated our Viral Vectors 101 eBook and blog posts. The lessons learned from this process have shaped and supported our growth and development. We’ve shared what we’ve learned, through manuscripts (Ersing et al., 2023; Haery et al., 2019; McDade et al., 2016), video, educational resources, and collaborations, and will continue to do so, in the hopes that these findings can help others as much as they've helped us. 

Interested in providing feedback on AAVs at Addgene? Suggest a viral vector for our service; leave a comment; or take our AAV survey.

Learn more about GCaMPs and other tools developed through the GENIE project at HHMI’s Janelia Research Campus on the GENIE website. 

---

 Bibliography 

Ersing, I., Rego, M., Wang, C., Zhang, Y., DeMaio, K. H., Tillgren, M., Fava, A., Clouse, G., Patrick, M., Guerin, K., & Fan, M. (2023). Quality control for Adeno-associated viral vector production. Neuromethods, 195, 77–101. https://doi.org/10.1007/978-1-0716-2918-5_5

Haery, L., Deverman, B. E., Matho, K. S., Cetin, A., Woodard, K., Cepko, C., Guerin, K. I., Rego, M. A., Ersing, I., Bachle, S. M., Kamens, J., & Fan, M. (2019). Adeno-Associated Virus Technologies and Methods for Targeted Neuronal Manipulation. Frontiers in Neuroanatomy, 13. https://www.frontiersin.org/articles/10.3389/fnana.2019.00093

McDade, J. R., Waxmonsky, N. C., Swanson, L. E., & Fan, M. (2016). Practical Considerations for Using Pooled Lentiviral CRISPR Libraries. Current Protocols in Molecular Biology, 115, 31.5.1-31.5.13. https://doi.org/10.1002/cpmb.8