It’s possible to use advanced analytics to predict biopharmaceutical failures, allowing corrective action to be taken before batches are lost or quality is degraded. A Seeq expert explains.
Joe Reckamp, Seeq
[Chromatography image courtesy of Seeq]
The biopharmaceutical industry uses chromatography processes extensively. Along with filtration, they are one of the most important separation and purification unit operations. These two processes affect quality and consistency significantly.
Chromatography is the principal purification unit operation in biologics synthesis, and it requires precise monitoring of column integrity and efficiency. Chromatography columns consist of packed resin or media used to separate solution components based on chemical or physical properties such as size, charge, hydrophobicity, or affinity. As the chromatography column is cycled, degradation of the resin ligand can decrease the consistency and effectiveness of the purification process.
The condition, degradation, and efficacy of the chromatography column is most commonly evaluated using transition analysis [Ref. 1-3]. Step change transitions in the column input solution, measured by conductivity or UV, are evaluated and reported as key performance indicators (KPIs). The standard KPIs used in the industry are height equivalent of a theoretical plate (HETP) and asymmetry [Ref. 1-2]. Separation efficiency of the chromatography column is defined by HETP, while asymmetry evaluates the normality of the peak to indicate the amount of peak fronting or tailing, each of which can result in reduced product quality and purity. These KPIs are helpful for monitoring column integrity and efficiency, as higher HETP and abnormal asymmetry values, or deviations from a value of 1.0, indicate resin degradation and potential batch failure.
Leading challenges with transition analysis calculations include data quality issues, the complexity of the differential equations used in the HETP calculation as these require moment analysis, and the time needed to conduct the analysis and alert the operator of a column failure. Traditionally, these complex calculations have been performed offline with mathematical software, which results in excessive time to insight. Operators are thus often reacting to column failure, rather than predicting or monitoring for imminent problems.
Seeq, an advanced analytics application used by process engineers and scientists, can accelerate the transition analysis calculation process by directly connecting to the data of interest and performing the required complex calculations (Figure 1).
Figure 1: Seeq connects directly to data to provide near real-time data analytics, along with reporting and dashboarding, to drive decisions. [Image courtesy of Seeq]
Below are three examples showcasing how Seeq software helped cleanse input conductivity data, conduct transition analysis, create online dashboards to reduce time to insight, develop prediction models to trend the KPIs over time, and apply new, more consistent, approaches to transition analysis involving TransWidth and DirectAf [Ref. 3]. TransWidth, an alternative to HETP, is a measure of the separation efficiency or resolution of the peak, while DirectAf indicates the normality or tailing of the peak and can be used in place of asymmetry [Ref. 3].
1. How Seeq software can help with data cleansing
Data quality is critical in advanced analytics such as transition analysis. There’s a need to collect data at a sufficient frequency. There’s the requirement of cleansing to remove outliers — such as those produced when a sensor is not in use — or filtering to smooth noise in the signal. Transition analysis requires differential equations to quantify the change in conductivity with respect to volume over the entirety of the transition period. HETP and asymmetry results can vary significantly depending on the frequency of the data and the type of interpolation used between the data points, triggering false column failure alerts or missing actual column failures if data is not handled properly.
It’s possible to use Seeq data cleansing tools on historized column data to remove outliers, focus calculations on only the transition time periods, and smooth the data. These three techniques help make the HETP calculations more consistent by isolating relevant data and removing noise (Figure 2), which enables engineers and scientists to increase precision and rigor in transition analysis calculations.
Figure 2: Data cleansing using Seeq’s filtering tools improves chromatography data for more precise transition analysis. [Image courtesy of Seeq]
Multiple pharmaceutical companies have used Seeq to remove outliers and cleanse conductivity data prior to transition analysis, dramatically improving the effectiveness of HETP trends over time. The KPIs monitored in transition analysis are expected to have low variation while the column is intact, but noise or inappropriate data sampling frequency can result in significant changes in the transition analysis KPIs, which may falsely trigger column failures. Data cleansing is used to enable effective monitoring of column health by reducing the dependence of the KPIs on the data sampling parameters.
2. Creating an online dashboard
The ultimate goal of transition analysis is to have the calculations performed online automatically with minimal delay between data collection and access to results of the calculations. The results can then be shared with operators, who can then prescribe actions such as regenerating the resin or repacking the column when the HETP or asymmetry values are out of specification to avoid quality deviations and lost batches. With Seeq, it’s possible to achieve these goals by performing the transition analysis calculations in Seeq Workbench and displaying the results in an online dashboard created using Seeq Organizer.
First, after the data have been cleansed, the Seeq Profile Search tools can be used to find phase transitions, rapidly detecting all similar transitions in the conductivity signal. For example, note in Figure 3 that Seeq Profile Search properly identifies the transition events and omits other similar shapes in the signal. Some systems, such as in multicolumn continuous chromatography, contain data from multiple columns. Changes in other signals, such as the differential pressures across each column, can be used to associate transitions with their respective chromatography column in order to track integrity and efficiency for each column. However, a similar approach can be used for single column chromatography if the column is cycled numerous times.
Both HETP and asymmetry calculations can be performed using the transition periods for each column. HETP is commonly calculated using moment analysis to describe the change in conductivity over column volume during each transition period. Asymmetry is calculated by comparing the change in column volume between the left and right sides of the conductivity transition peak. These calculations can be performed using the Formula tool in Seeq.
While transition analysis can be performed in other calculation programs, the complexity of the calculations often results in significant delays between when the data is generated and when the results of the analysis are complete. In addition, the calculation process entails extracting the data from historians, inserting the data into a calculation program, and communicating the results as separate steps. Any delays in this process can result in missed column failures leading to lost or reworked batches, reducing product yield and resulting in millions of dollars in lost product.
Seeq streamlines this workflow by connecting directly to data in all leading historians, performing the calculations automatically as new data are collected, and communicating the results through auto-updating dashboards (Figure 3).
Figure 3: Online dashboards can be developed to monitor multiple chromatography columns in near real-time. [Image courtesy of Seeq]
Pharmaceutical companies have utilized Seeq Organizer to successfully monitor chromatography column health by creating online production dashboards to monitor HETP and asymmetry. Using Seeq, these companies have realized savings up to 10 hours per week per unit when performing transition analysis calculations, with further financial benefits gained by applying predictive maintenance to limit unplanned downtime and subsequent decreases in batches produced.
3. Building predictive models for transition analysis
Over time, chromatography resin degrades, becoming less efficient at separating biologics and resulting in increasing HETP and asymmetry values as the resin reaches the end of its life. Tracking HETP and asymmetry over time or batches enables corrective action prior to column failure.
Seeq Workbench enables subject matter experts to access process data, define the equations for transition analysis, and build models to forecast predicted HETP and asymmetry values. The Prediction tool within Seeq can create multivariate predictive models using principal component analysis or other regression algorithms to predict HETP and asymmetry values based on current utilization of the chromatography columns (Figure 4).
Figure 4: Transition analysis HETP and asymmetry values can trend toward column failure over extended periods of time. Models can be applied in Seeq to predict the amount of time until column failure. [Image courtesy of Seeq]
Predictive models can be used to forecast an appropriate maintenance window prior to column failure. Performing online transition analysis with predictive modeling in Seeq can thus reduce downtime, product quality deviations, and lost batches.
4. TransWidth and DirectAf transition analysis
Alternatively, scientists and engineers can use the new TransWidth and DirectAf approach for transition analysis [Ref. 3]. TransWidth, an alternative to HETP, is a measure of the separation efficiency or resolution of the peak, while DirectAf indicates the normality or tailing of the peak and can be used in place of asymmetry [Ref. 3]. TransWidth and DirectAf have been shown to be less influenced by noise in the data and thus may be more robust metrics for detecting column integrity and efficiency.
Just–Evotec Biologics, Inc. has used Seeq to calculate TransWidth and DirectAf for chromatography column monitoring. The reduction in noise using the TransWidth KPI is evidenced in Figure 5, in which the blue TransWidth values vary less than the HETP values calculated using the traditional moment-based approach. Figure 5 also shows evidence of column collapse detected by both TransWidth and HETP.
Figure 5: Just – Evotec Biologics, Inc. has used TransWidth and DirectAf calculations in Seeq to monitor column integrity over time for their chromatography columns. [Image courtesy of Seeq]
These techniques have allowed Just–Evotec Biologics to improve their production processes.
Conclusion
Transition analysis can be an effective method for tracking chromatography column efficiency and predicting failure. However, current challenges with the time to insight, the complexity of the calculations, and the quality of the data have limited the ability of biopharmaceutical companies to operationalize transition analysis calculations, and to then accurately identify and predict column failures quickly to reduce lost product. These issues are compounded in continuous, multicolumn production systems running simultaneously over extended periods.
Seeq enables rapid time to insight for transition analysis by connecting directly to data sources and automatically performing calculations as new data are generated for both single- and multi-column systems. Seeq has worked with a number of pharmaceutical companies to enable monitoring and prediction of chromatography column health via built-in data cleansing tools and the ability to create online KPI monitoring dashboards, develop predictive models, and perform alternate column integrity calculation methods.
Joseph Reckamp is an analytics engineer at Seeq (Seattle), specializing in the pharmaceutical industry. He enjoys working with engineers across manufacturing industries to improve processes and realize value through the use of process data analytics. He received his B.S. and M.S. in chemical engineering from Villanova University and has worked in the pharmaceutical industry throughout his career, including stints in in research and development with GlaxoSmithKline and production with Evonik.
References
[1] B. McCoy & M. Goto (1994). Continuous Mixture Model of Chromatographic Separation. Chem. Eng. Sci. 49 (14) 2351-2357.
[2] T.M. Larson, J. Davis, H. Lam, & J. Cacia (2003). Use of Process Data to Assess Chromatographic Performance in Production-Scale Protein Purification Columns. Biotechnology progress. 19. 485-92. 10.1021/bp025639g.
[3] Y. Cui, Z. Huang, & J. Prior (2018). Using Direct Transition Analysis in Chromatography. BioPharm International 31 (1) 2018.