CryoTEM on the USP Certificate: A Milestone for AAV Capsid Characterization

CryoTEM on the USP Certificate: A Milestone for AAV Capsid Characterization

In May 2026, the United States Pharmacopeia added a section that the AAV characterization community has been waiting for. Both of USP's AAV8 reference standards — Full Capsids (Cat. No. 1000302, Lot F238Q0) and Empty Capsids (Cat. No. 1000301, Lot F238K0) — now carry CryoTEM data on their certificates. The full-capsids certificate moved to v02 on 14-MAY-2026, and the empty-capsids certificate to v01 on 13-MAY-2026, in both cases with the same reason for change: Addition of CryoTEM.

For those of us who run CryoTEM as a characterization service for gene therapy developers, this is a moment worth marking.

Why the full/empty ratio keeps coming back

The fraction of AAV capsids carrying the intended genome versus those that are empty, partially filled, or overfull is one of the most consequential critical quality attributes in recombinant AAV manufacturing. It drives potency calculations, informs dose, and shapes the immunogenicity profile of the final product. Regulators expect it to be quantified, and they increasingly expect it to be quantified by more than one orthogonal method.

The challenge is that no single technique gives a clean, ground-truth answer. Every method has assumptions baked into its physics: SEC-MALS infers composition from light scattering and refractive index, AUC from sedimentation behavior, A260/A280 UV ratios from bulk absorbance, mass photometry and CDMS from single-particle mass distributions. Each is rigorous; none is direct.

What makes CryoTEM different, and why USP's move matters

CryoTEM is the only commonly used method that classifies capsids by looking at them. A vitrified specimen on a Glacios at 200 kV gives you a micrograph in which full, intermediate, empty, and overfull particles are distinguishable by contrast and morphology. Each particle gets counted individually. There's no bulk-property inference and no calibration curve standing between the measurement and the answer.

That directness is exactly why it pairs so well with the inferential methods already on USP's certificates. Cross-referencing the full-capsids certificate makes the case better than any argument:

Six independent techniques, six values within a roughly 18-point window, all converging on the same biological truth. CryoTEM sits comfortably in the middle of the pack — exactly where a well-executed direct count should.

The practical implication for QC labs

Until now, a lab running CryoTEM on its AAV process samples had no pharmacopeial anchor to validate against. You could agree with your AUC. You could agree with your mass photometry. But there was no reference standard with an officially documented CryoTEM value to demonstrate that your imaging, segmentation, and classification workflow was producing results consistent with a recognized authority.

That anchor now exists. Labs can run USP 1000302 and 1000301 through their CryoTEM pipelines and check their full / intermediate / empty / overfull distributions against the published values (89 / 3 / 5 / 3 for full capsids; 1 / 0 / 98 / 0 for empty capsids). For method qualification packages, regulatory submissions, and inter-lab comparability studies, that's a meaningful upgrade.

Where we sit

We've been running CryoTEM as a service for AAV developers since well before USP formalized it on a certificate, and our workflow — vitrification, low-dose imaging, automated particle picking, contrast-based classification — is structurally identical to what produced the data USP has now published. We've already benchmarked our pipeline against both reference standards and will be publishing the results shortly.

For developers building out their CMC story, this is a good moment to revisit how full/empty is being measured in your program. If your current orthogonal panel doesn't include a direct imaging method, the gap is more conspicuous than it was a month ago. If it does, you now have a way to prove it works.

We're happy to talk through either case.

‍