Data Collection and Experimental Input

Expertise as a Starting Point

Effective modelling in pharmaceutical development relies fundamentally on the quality and relevance of the underlying data. A first advantage in Elegent’s data collection strategy is the combination of material science expertise and process understanding.

The experimental work is guided by in-depth knowledge of:

• The diverse physical natures of pharmaceutical powders
• The deformation, flow, and friction mechanisms governing solid processing
• The interaction between material properties and equipment behavior

This understanding ensures that data collection is not performed in a purely exploratory manner, but is instead targeted toward measurements that are mechanistically meaningful and directly relevant for modelling and decision-making.

Structured Use of Existing Data

Elegent builds upon a substantial and continuously expanding experimental database. Beyond its direct use for model training and validation, this dataset is actively exploited to guide further data generation.

By applying data decomposition techniques to existing measurements, it becomes possible to:

• Identify underrepresented regions in formulation or material space
• Detect combinations of powder properties that are expected to exhibit distinctive behavior
• Propose new, informative formulations for experimental evaluation

In this way, historical data is not treated as static input, but as a dynamic resource that informs and refines future experimental design.

Model-Guided Data Collection

Predictive models play a central role in ensuring efficient data collection. In particular, Elegent’s in-house models for:

• Mixed powder properties
• Critical quality attributes (CQAs)

are used to prioritize experiments that are most likely to add value. Model predictions help:

• Focus experiments on regions of high uncertainty or high sensitivity
• Avoid redundant measurements in well-characterized areas
• Accelerate learning across formulation and process space

This iterative interaction between experimentation and modelling supports a data-efficient strategy, where each new experiment is selected to maximally improve understanding and predictive capability.

Summary

Elegent’s data collection capabilities are characterized by the integration of experimental expertise, structured use of existing data, and model-guided experimentation. This approach ensures that data generation remains purposeful, scalable, and closely aligned with the needs of formulation development and predictive modelling.