
Researchers at Purdue University, with a co-author from Eli Lilly and Company, report an analytical method for adeno-associated virus (AAV) manufacturing that uses quantum cascade laser-based mid-infrared spectroscopy (QCL-MIDIR). The paper appeared online on 3 Nov 2026 in Analytical Chemistry. This note covers the abstract only, because the full text is behind a subscription.
The problem the paper addresses
The abstract says AAV production remains challenging because of incorrectly filled particles (capsids containing partial genome information), residual host cell proteins and residual host cell DNA. In the system studied, the host cells are Spodoptera frugiperda (Sf9) cells infected with baculovirus. The authors write that empty and partially filled capsids must be monitored closely, and that current analytical methods are labor-intensive, costly and slow, with low throughput.
What was done
The authors demonstrate on-line applications of QCL-MIDIR to characterize AAV9 particles and to estimate two quantities: the capsid concentration and the percentage of empty or non-empty capsids. They built partial least squares regression (PLSR) models, a standard statistical method that links spectra to measured values, to predict both quantities.
The reported detection limits
- Concentration: the model can detect AAV9 particles with a minimum limit of detection (the lowest concentration it can pick up) of 7.39 × 1012 capsids per mL.
- Fill state: the model can detect down to 15.02% non-empty capsids.
The abstract does not state prediction errors, sample numbers or how the models compare with reference methods. It also does not say which process step or sample type the on-line measurements came from beyond the cell culture lysate context. Those details would be in the full text.


