ACD Labs Logo
MENU

HTC-17

May 18-20, 2022
The Aula, Ghent University, Ghent, Belgium

Presentation Schedule

Prediction and simulation tools for more efficient method development
Shahriar Jahanbakht, Andrew Illsley, James Hogbin, Charis Lam, Andrey Vazhentsev, Roman Yurov

View Abstract

Prediction and simulation tools for more efficient method development
Shahriar Jahanbakht, Andrew Illsley, James Hogbin, Charis Lam, Andrey Vazhentsev, Roman Yurov

Developing quality methods requires an understanding of the entire chromatographic problem. How do analytes physicochemical characteristics affect their retention? How do controllable factors, such as gradient, temperature, and pH, affect desired measures, such as resolution? While repeated experiments supply that knowledge, they also consume significant time, manpower, and resources.

Software can help chromatographers understand the separation problem in less time and with fewer experiments. Here, we present ACD/Method Selection Suite, a method development assistant that includes tools for predicting physicochemical properties and simulating separations.

The prediction tool was used to predict the pKa of dazatinib and two impurities, and their logD at different pHs. Ranges were identified where the dominant ionic form was robust to changes in pH. These areas are particularly suitable for initial pH screening experiments.

The simulator tool was used to optimize the separation of three compounds in 3D mode. A 2x2x3 matrix of experimental conditions was designed to vary gradient, temperature, and pH, and a 3D rotating-cube model was produced to visualize predicted resolution at each point. The 3D map also marked areas of poor robustness. While a few of the initial conditions failed to resolve all three compounds, the simulator suggested several conditions that produced good resolution.

The user retains control over the simulator and can adjust settings based upon their chromatographic knowledge. For example, modelling equations can be customized to improve model accuracy, and the success criteria can be modified.

This work demonstrates two separate tools to help chromatographers understand factors governing analyte separation. The use of such software can reduce the time and number of experiments needed to develop a separation method. Such tools could be further extended by combining them with other useful chromatographic algorithms like retention time prediction, to optimize separations containing new but related analytes with a minimal number of experiments.