Applying Unified Laboratory Intelligence in a High-Throughput, Multi-Technique Environment
Gathering all analytical data from different techniques for the same sample isn’t always an easy and routine task. This problem is amplified in high-throughput environments based on sheer volume alone. Review and analysis of information can be time consuming, leading to delays in decision-making that have detrimental effects on productivity and the speed of project completion. High-throughput analytical chemistry is applied in many environments including formulations and quality control. This article discusses high-throughput analytical chemistry in the context of preformulation in the pharmaceutical industry.
Preformulation studies occur between the late phases of discovery and the early phases of development, sometimes referred to as the interface between a drug substance and a drug product. To design an optimum drug delivery system, an understanding of solid-state properties is necessary to evaluate drug candidates for their development risks and evaluate whether this target drug should move forward in the process.
The two primary goals of the preformulation scientist are to find all possible salt and polymorph forms of the active pharmaceutical ingredient (API) and to identify the best form to move forward with, all in the shortest time possible. Finding all possible forms is critical for protecting their patent, while finding the best form involves many different studies to understand stability, compatibility with excipients, bioavailability, solubility and other solid-state properties. The pressure is on to get answers quickly, because once the patent process is underway time is money. Because the understanding of how salts and polymorphs are created is limited, the common approach is to run multiple analytical experiments to explore many possible combinations of factors to create salts and discover what polymorph forms can exist. Analytical techniques typically used in preformulation studies include primarily XRPD along with, DSC, TGA, HPLC, NMR and other studies.
Manage unified analytical information
In today’s laboratories, analysts and groups deal with the vast volume of disparate data in a number of ways. They either share information from various sources and techniques on paper/electronic PDFs and carry out manual review and analysis, or develop in-house software to help manage the exercise. While in the best-case scenario a digital image of the data may be available to make data sharing easier, the inability to re-process or overlay multiple pieces of data limits efficiency. In an ideal situation, analysts would bring information into a uniform shared database as it is created. With all the relevant data in a searchable, electronic environment, review, analysis and decision making in the identification of salts and polymorphs maximizes the scientist’s throughput.
Unified Laboratory Intelligence (ULI) is a technology framework to collect and unify chemical, structural and analytical data, which combines analytical content with chemical context. It offers homogenization of disparate data from different analytical techniques and labs/groups, including data from other informatics systems such as ELNs or LIMS, onto a single platform. This “live” data can be shared across groups, departments and sites; and when searched, retrieved and applied (or re-applied), enables preformulation scientists to make faster decisions. Having all the data in one environment with tools to review and analyze it further improves productivity. Moreover, automating data aggregation and processing procedures delivers further benefits—less manual intervention in gathering information, data availability in real time and reduction of errors in matching different data types for the same sample.
Generate knowledge from information
Once the information from related experiments is brought together, scientists analyze it to determine which experiments form unique salts and polymorphs by comparing all of the XRPD spectra (diffractograms) with each other. Since several hundred XRPD experiments may be run for a single project, examining this data can be a daunting task. Cluster analysis helps the scientist group similar spectra together, indicating unique forms. As the number of experiments increases, a manual clustering process becomes more time consuming and difficult. Clustering algorithms are useful, giving the user a starting point for what data to examine, and to help extract knowledge from the data.
Review and decision making
A number of problems are encountered when clustering XRPD data for salts and polymorphs, the most important being clustering accuracy. Because clustering algorithms are inexact, data must be reviewed by an experienced analyst to determine the accuracy of the clustering model. A ULI-applied approach provides dynamic visualization of the data and delivers flexibility in data manipulation to fit different workflows and data review methods that may be employed within the group.
Comparing multiple spectra is one of the most important aspects of the data review process and, as mentioned earlier, paper printouts and electronic spectra are limiting since spectra can’t easily be overlaid for direct comparison. In these formats, the information is essentially frozen as dead data. In a “live” spectra environment, data can be overlaid and manipulated enabling analysts to make confident decisions based on knowledge extracted from analytical information. The process of selecting which spectra to compare is also important; so having several methods to select, analyze and compare data is imperative. A single representative spectrum may be set for comparison with a group of spectra, the group may be compared to an average spectrum, multiple spectra may be selected or a “nearest neighbor” table can be used to help guide the movement of spectra between clusters.
Create intelligence to gain insights
Sharing the knowledge accumulated daily by scientists involved in preformulation studies can provide important information in later processes—whether examining a future production problem or providing guidance for re-formulation of the product to extend a patent. The live data may also be leveraged to accelerate future projects and experiments, empowering scientists to make better estimations of what experiments may generate salts or polymorphs for a given class of materials.
Employing a ULI technology framework to capture and retain live multi-technique analytical and chemical data in a uniform environment streamlines data analysis and management for scientists. The decision-making process is accelerated through dynamic data visualization, advanced algorithmic tools (aiding analysis) and access to legacy data. In turn, organizations benefit from centralized analytical and chemical knowledge management which leads to enhanced productivity. The process of gathering and analyzing high-throughput analytical data is facilitated from the initial gathering of data, to final report compilation. While ULI will provide value to any analytical laboratory, it seems imperative for scientists that work in high-throughput analytical environments to adopt these technologies that will ease the burden of tedious tasks and improve efficiency and ROI.