March 22, 2024
by Sarah Srokosz, Marketing Communications Specialist, ACD/Labs
A Digital Twin for Analysis Workflows
As one of the leading manufacturers and distributors of certified reference materials in the world, MilliporeSigma, the U.S. and Canada Life Science business of Merck KGaA, Darmstadt, Germany has recognized the growing need for digital solutions in this area. They have taken the first step towards meeting the analytical chemist’s need for digitalization with the release of their digital reference material platform, ChemisTwin™.
ChemisTwin™ is a cloud-based platform which provides scientists in quality control labs with an automated structure verification (ASV) tool tailored to their workflows. By combining their expertise in manufacturing high quality reference materials with the industry-leading ASV and spectral prediction capabilities of ACD/NMR Workbook Suite, this new tool from MilliporeSigma is set to revolutionize the way scientists approach analytical testing.
The Spotify of Analytical Chemistry?
In an episode of ACD/Labs’ podcast, The Analytical Wavelength, we asked ChemisTwin™ Project Lead Coralie Leonard about what motivated MilliporeSigma to embark on such a drastic shift in their approach to their more than 25,000 reference materials. She explained that feedback from their physical reference material customers told them that their industries need:
- Greener alternatives to common processes
- Increased efficiency in routine tasks
- Improved access to reliable reference materials
To address these, she and her colleagues asked,
“Could we find a really original way within our area of expertise of producing physical reference material, to use this expertise to solve some of those problems for our chemists?”
MilliporeSigma know that when they send a customer a physical reference material, what the scientist ultimately cares about is the data. This led them to the idea that they could send the data directly, eliminating the need for production, shipping, and disposal of the chemical. Additionally, unlike a physical sample, the data can be made immediately available. Coralie compares the philosophy behind ChemisTwin™ to the music industry’s transition from physical media like CDs to digital streaming platforms.
And thus, ChemisTwin™ began to take shape—a digital twin of MilliporeSigma’s entire catalog of >25,000 physical reference material on a cloud platform where scientists can upload their own data to compare with high-quality data of the reference material. This ultimately led MilliporeSigma’s team to coin the term digital reference material (dRM)—a dataset detailing the properties of a physical reference material, such as spectral information, together with information like their CAS number, and other analytical data.
Leveraging ACD/Labs’ NMR Data Analysis Expertise
With a catalogue of this size, Coralie’s team knew that they would have to be strategic about where to start this digital transformation in order to realize the potential benefits they saw in it. For the first release, they decided to focus on routine-type structure verification workflows using NMR data. The repeated nature of such analyses means they would benefit the most from efficiency gains with ChemisTwin™. They chose NMR data over that of other analytical techniques, as its continued development, versatility, and unique characteristics lead them to believe it is a “technique of the future” that can impact a wide variety of industries.
NMR data was the first choice for ChemisTwin™ because it is:
- Already widely employed for a diverse range of applications
- Sustainably produced—uses minimal solvent, non-destructive
- Highly reproducible—consistency and reliability across different instruments (especially important for their targeted audiences)
- Suitable for a wide variety of sample types (organic compounds, inorganic compounds, small molecules, complex mixtures)
- Indicative of the arrangement of atoms in a compound
- Qualitative and quantitative
With their focus set on NMR, MilliporeSigma chose the industry-leading processing, prediction, and analysis capabilities of NMR Workbook Suite to play a pivotal role in this project, both in the creation of the dRMs and the analysis of a user’s sample data.
The Making of a Digital Reference Material (dRM)
The process of making a dRM is different in many ways from that of a physical reference material, though Coralie and her colleague, qNMR R&D Scientist Albert Farré Pérez, highlighted that they share the same three key steps: production, inspection, and review. For a digital reference material, the production step consists of an analyst acquiring a spectrum of the specified material. They ensure that dRMs are traceable to the equivalent well-characterized physical reference material. When available, reference materials certified under ISO 17034 are used.
In order to make this dataset into a dRM, they import it, along with the corresponding structure, to their customized version of NMR Workbook Suite, that is built-in to the ChemisTwin™ platform. They perform a preliminary qualitative assessment of the raw data to ensure that the correct solvent is used, there are no impurities, there is good solubility of the compound, and there is no large overlap between sample and solvent peaks. Subsequently, the spectrum is processed, and peaks are assigned to the corresponding protons in the structure. This production step is followed by an inspection, and then a review step under the “six eyes” principle of ISO 9001. The dataset is then used to train ChemisTwin’s NMR Workbook Suite prediction algorithm and the dRM is released.
How it Works
When scientists upload their raw data to ChemisTwin™, it uses this customized version of NMR Workbook Suite to analyze and compare their data and the applicable dRM(s) in a workflow that is similar to batch automated structure verification.
ChemisTwin™ automatically extracts the metadata from the sample data. From there, NMR Workbook Suite’s prediction algorithm generates a predicted spectrum that accounts for differences in experimental settings (such as solvent and field strength) between the certified data set in a dRM and a user’s sample data. It then compares the experimental spectrum to this predicted one, evaluates the match between them, and calculates the Match Factor (MF). If the quantitative workflow is chosen, NMR Workbook Suite’s qNMR algorithms will calculate the amount present in the sample. ChemisTwin™ then compiles these analysis results in a report for the scientist to review and send for further approval, if required.
The Promise of Automated Analysis with On-Demand Data
First and foremost, ChemisTwin™ positions scientists and their organizations to save an unprecedented amount of time and resources by eliminating or automating nearly every step in the applicable workflow.
dRMs eliminate the need for scientists to procure, prepare, acquire data of, and dispose of physical reference materials. The time and resource savings from this aspect alone can quickly add up due to the prevalence of these analyses in the pharmaceutical and other related industries. Together with the automated data analysis mentioned above, the scientists’ workload is significantly reduced.
In addition to this, ChemisTwin™ can provide further benefits that are realized at an organizational level. Having all analyses done in the same way, on a single platform, using the same dRM creates harmonization across an organization’s data, regardless of the location, instrument, or analyst involved. Additionally, by eliminating the need for highly trained NMR scientists in routine analyses, the platform has the potential to make NMR data’s valuable insights more accessible and allows experts to focus on more challenging tasks.
What’s Ahead for ChemisTwin™?
Currently, ChemisTwin™ can perform targeted and non-targeted qualitative analysis of fairly pure samples based on the approximately 2000 available dRMs. These have been created through the process mentioned previously, which ensures the traceability back to the physical reference material. Additionally, there are 57,000 theoretical dRMs also accessible in the platform which have been generated using the training algorithm. Customers can utilize them as an initial reference for conducting analysis until a dRM is created. Customers also have the option to request specific dRMs through the portal and these requests are being prioritized in the production process.
Navigating Regulatory Acceptance
Because the concept of dRMs is entirely new to the industry, Coralie explained that analyses using dRMs are not automatically accepted by regulatory authorities in the same way as when you use a certified reference material or equivalent. However, she and her team are actively working with their accreditation body, some regulatory institutes, and authorities to build a framework where these can be more automatically recognized by regulatory authorities.
The Envisioned Evolution of ChemisTwin™
As for what’s in the platform’s future, the ChemisTwin™ team is continually releasing more dRMs and training the algorithms to further improve the accuracy of the theoretical dRMs. Additionally, tools for qualitative analysis of mixtures and quantitation are currently in development. They are also working towards incorporating data from other analytical techniques to create a complete digital twin of the entire MilliporeSigma reference material catalog.
Step into the Future of Analytical Testing
The convergence of MilliporeSigma’s expertise in physical reference materials and ACD/Labs’ industry-leading analytical data processing and interpretation technology has yielded a platform that not only addresses current challenges but also paves the way for a more accessible and efficient future in analytical testing.
Learn more by listening to the full conversation with Coralie and Albert in this episode of The Analytical Wavelength. You can also watch presentations on ChemisTwin™ from our previous online events by both Coralie and Albert.
ChemisTwin™ is available now and you can try it for free for a limited time.
About the Author
