Book a Live Demo

Chromatography and Transition Analysis

Challenge

Chromatography is a key unit operation for producing a high-quality pharmaceutical. A biotechnology company specializing in the design and manufacture of proteins was struggling with the incredibly time-consuming process of comparing the quality and yield of batches based on the chromatography peaks in Excel. In addition, chromatography columns degrade over time and require integrity and efficiency monitoring to predict column failure to avoid a potential loss of a batch.

Solution

To capture quality and yield of chromatography peaks, the peaks were overlaid in Seeq with the mass yield calculated from the UV absorbance and flow rate signals. Transition analysis was performed using both the Height Equivalent of a Theoretical Plate (HETP) and Asymmetry methods as well as the TransWidth and DirectAf methods to track column integrity and efficiency of the separation. All the analyses were then shared throughout the organization with auto-updating dashboards to track the column integrity, process yield, and quality in near real-time.

Results

Chromatography peak calculations for quality and yield save an average of 15 hours per week. Transition Analysis calculations enable monitoring of column integrity, which could save multiple batches each year, each potentially saving millions of dollars.

Data Sources

  • Emerson DeltaV
  • OSIsoft PI
  • CSV Files

Data Cleansing

  • Agile Filter performed on conductivity data for transition analysis
  • Calculated normalized conductivity traces for TransWidth and DirectAf transition analysis

Calculations and Conditions

  • Profile Search used to identify chromatography peaks and transitions
  • Calculated mass of chromatography peaks
  • Calculated transition analysis via multiple methods following literature calculations

Reporting and Collaboration

  • Auto-updating dashboards in Seeq Organizer
  • Chromatography overlays tracked in daily meetings
  • Column integrity tracked via transition analysis, identifying potential column failure to operators in near real-time