• Sponsored Content

BPI Staff

October 15, 2015

4 Min Read

BPI-AtE-logo_1-300x94.jpg with Thomas Flouquet (Novasep)

The first simulation and optimization software for chromatography process development in biopharmaceutical purification, the BioSC Predict program from Novasep received an 2015 Achema award for Pharmaceutical Engineering. Driven by user objectives, it can define practical constraints and desired strategies. Optimization algorithms provide the most adapted operational conditions for a BioSC system. In a webcast on 17 September 2015, Novasep’s Thomas Flouquet demonstrated the software’s functionality and customizability. He presented optimized recipes taking into account parameters such as process type, feed titer, load velocity, number of columns, and resin bed height. Below is a summary of his presentation. You can find more detail and the full slide presentation online.

Flouquet’s Presentation
BioSC Predict simulation software optimizes the protein capture step with affinity, ion-exchange, and hydrophobic solvents. It maximizes solvent use to achieve a user’s objectives related to time-savings and productivity. It also sets parameters to program a laboratory or pilot-scale BioSC continuous chromatography unit.

The first stage of a complete simulation is “My Batch Process.” It begins with batch-process conditions and performance information that users enter. For each step — load, three different washes, elution, regeneration, and equilibration — a user provides the necessary information to fit the software. The program then generates results regarding that batch process (e.g., productivity and how fast a gram of product is purified) and a simulated process (optimized batch, optimized parallel batch, or optimized continuous chromatography).

The second stage is breakthrough-curve modelization, in which the software defines key factors, thermodynamics, and kinetics parameters for the same feed and solvent. Ideally I run three different breakthrough curves at three different velocities (at least one curve reaching 80% breakthrough), with the same column and solvent, after providing information on the feed titer, column diameter, and bed height. I could import data (e.g., from Microsoft Excel) and then click on “modelize.” Now I have three different breakthrough curves at three different velocities, and the software gives me the static binding capacity (SBC, the maximum binding capacity of my solvent for my feed). It has defined the thermodynamics and kinetics parameters of my feed and solvent to simulate an optimized recipe.

Simulation is the third stage. If I want an optimized recipe, I start with BioSC process parameters. Then I provide ranges for four optimization parameters (the range of freedom I give the software to optimize my recipe): the numbers of columns, bed height, loading velocity, and loading capacity. The latter is how much I want to saturate the solvents. The software generates a table describing the steps of “my batch process”: load, three different washes, elution, regeneration, and equilibration. For each step, I have the linear velocity and bed volume from the first stage of the simulation (the batch process).

So I’m telling the software to optimize velocity and design a recipe with two to six columns. For column lengths, you can use from 5–15 cm, all with the same bed height. But it’s up to the software to define the optimized bed height, loading velocity, and loading capacity. Once those optimization parameters have been entered, I will click on “optimize.” To optimize the loading steps, all the other steps (washes, elution, regeneration, and equilibration) will remain the same. The software also provides information about period and cycle time. In one cycle, all columns have been loaded and processed.

For maximized productivity, you choose the time-saving scenario option: asking the software to design a recipe to process as fast as possible. But there are other scenarios: sorbent-saving and an option to balance time and sorbent savings. Once you are happy with what has been designed by the simulation software, you can have your BioSC unit run that recipe.

For maximized sorbent material saving, we have another loading step. This is a bit slower than the previous scenario, but much more saturated. And a balance scenario reaches a saturation between the two using the same columns. You can compare scenarios for economical performance.

Industrial-Scale Example: Consider that you have to produce 5 kg within 12.8 hours. The software will tell you, for example, that you need four 35-cm columns with 5-cm bed height, with 19.3 L total feed. If you increase the quantity to be produced and reduce the production time, then you must increase column diameter and total volume amount of solvent required.

It’s not always beneficial to switch from batch to continuous processing. In such cases, the software will suggest parallel batches. BioSC Predict software generalizes continuous chromatography by using chromatography practice and knowledge to design the most consistent continuous chromatography or batch step to match a user’s expectation. Try BioSC Predict Lite software for free.

Watch the full presentation on-demand.

You May Also Like