BIOVIA Discoverant provides a powerful environment for descriptive and investigational analysis. The capabilities leverage BIOVIA’s flexible, on-demand data access and contextualization environment to support extensive analytical and statistical tools.
The data analysis modules enable process engineers to understand not only what is happening (descriptive analytics) but also why it is happening across the entire process (i.e., cause-and-effect analysis) and across operations occurring at different sites.
The modules were designed, developed and tested to be fully validatable.
Using BIOVIA’s data analysis capabilities, manufacturers can conduct:
Process and product trending with alerts for early warning
Tri-panel root cause investigations for relating events to process shifts
Correlation plot matrices for finding unexpected parameter relationships
Process Capability calculations and specifications setting
Comparison groups of batches to understand process variations
Basic and advanced univariate and multivariate statistical tools help scientists easily access and analyze everyday process data. For example, scientists can use Two-Sample t-tests to compare the average of two groups to find out if they are significantly different.
Principal Component Analysis (PCA)
Principal Component Analysis (PCA) can be used to transform a large data set with a large number of correlated variables into a smaller set of uncorrelated variables. These variables can be used in other types of analyses carried out both inside and outside the BIOVIA Discoverant application.
Plotting and linear regression modeling capabilities allow you to inspect and analyze stability study data in a flexible manner to calculate expiration estimates that conform to existing industry standards. Alternate linear regression modeling strategies can be used to pool production lots and obtain a more reduced model when it is appropriate to do so. Scientists can also use the Stability Out-of-Trend (OOT) alerting capabilities that allow detecting out-of-trend study results while stability studies are in progress. Different evaluation methods give you the ability to analyze stability parameter trends, either within batches or by comparison with trend estimates generated from a combined set of historical study results contained in a designated reference data set.
Curve Metrics® provides scientists with point-and-click tools needed to work comprehensively with continuous data and extract features on screen. Scientists can combine quantified features with other discrete and replicate data for analysis. With Advanced Profile Analysis (APA), scientists can analyze multiple time series parameters simultaneously to determine which of them in combination (and when) are the best predictors of the process outcome.
Visual Process Signatures® (VPS) provides animated visualizations of interactions among process parameters over time, or in relation to the value of a selected process parameter or process outcome. VPS can uncover otherwise hidden parameter relationships in the manufacturing process and identify parameters that warrant further analysis to improve process control.
Multi-Phase Analysis (MPA) provides enhanced ways to examine chromatography, fermentation and other multi-phase data. MPA enables automatic or point-and-click identification of the phases in continuous data. Once identified, phases can be visually and quantitatively compared individually or in groups. MPA can be used to more accurately determine the useful life of costly chromatography resins to avoid premature, or late, replacement, thereby minimizing the risk of impurities entering the product stream.
Scientists can compare data graphically by plotting a wide range of parameters displayed across multiple panels to analyze a single batch, multiple batches, or groups of batches. Summarization functions are enhanced to assist scientists in understanding the historical data. An easy point-and-click function allows scientists to overlay the progress of the current batch in real-time onto historical batches to assess performance against the "golden batch."
Quality Monitoring provides Statistical Process Control (SPC) for real-time process monitoring and automated alerts. Using a variety of control charts, scientists can plot and monitor a process visually against pre-determined limits. Process Capability functionality allows scientists to set multiple limits and see the predicted failure rates based on actual process performance.