Hybrid modeling of an ultracentrifugation process for separation of full and empty adeno-associated virus particles.

Ultracentrifugation is an attractive method for separating full and empty capsids, exploiting their density difference. Changes of the serotype/capsid, density of loading material, or the genetic information contained in the adeno-associated viruses (AAVs) require the adaptation of the harvesting parameters and the density gradient loaded onto the centrifuge. To streamline these adaptations, a mathematical model could support the design and testing of operating conditions.Here, hybrid models, which combine empirical functions with artificial neural networks, are proposed to describe the separation of full and empty capsids as a function of material and operational parameters, i.e., the harvest model. In addition, critical quality attributes are estimated by a quality model which is operating on top of the harvest model. The performance of these models was evaluated using test data and two additional blind runs. Also, a "what-if" analysis was conducted to investigate whether the models' predictions align with expectations.It is concluded that the models are sufficiently accurate to support the design of operating conditions, though the accuracy and applicability of the models can further be increased by training them on more specific data with higher variability.

A Simple Flow Classification Parameter Can Discriminate Rupture Status in Intracranial Aneurysms.

BACKGROUND: A simple dimensionless aneurysm number ($An$), which depends on geometry and flow pulsatility, was previously shown to distinguish the flow mode in intracranial aneurysms (IA): vortex mode with a dynamic vortex formation/evolution if $An > 1$, and cavity mode with a steady shear layer if $An < 1$. OBJECTIVE: To hypothesize that $An\ > \ 1$ can distinguish rupture status because vortex mode is associated with high oscillatory shear index, which, in turn, is statistically associated with rupture. METHODS: The above hypothesis is tested on a retrospective, consecutively collected database of 204 patient-specific IAs. The first 119 cases are assigned to training and the remainder to testing dataset. $An$ is calculated based on the pulsatility index (PI) approximated either from the literature or solving an optimization problem (denoted as$\ \widehat {PI}$). Student's t-test and logistic regression (LR) are used for hypothesis testing and data fitting, respectively. RESULTS: $An$ can significantly discriminate ruptured and unruptured status with 95% confidence level (P < .0001). $An$ (using PI) and $\widehat {An}$ (using $\widehat {PI}$) significantly predict the ruptured IAs (for training dataset $An\!:\ $AUC = 0.85, $\widehat {An}\!:\ $AUC = 0.90, and for testing dataset $An\!:\ $sensitivity = 94%, specificity = 33%, $\widehat {An}\!:\ $sensitivity = 93.1%, specificity = 52.85%). CONCLUSION: $An > 1$ predicts ruptured status. Unlike traditional hemodynamic parameters such as wall shear stress and oscillatory shear index, $An$ has a physical threshold of one (does not depend on statistical analysis) and does not require time-consuming flow simulations. Therefore, $An$ is a simple, practical discriminator of IA rupture status.