Innovation iBioG-Med Platform
Introduction
The iBioG-Med is a culture media optimization design platform developed by BioEngine independently through statistics, data mining and machine learning for the development of customized culture media formulations. The iBioG-Med platform allows rapid development of customized media for target cells, meeting customer expectations for cells/products as well as meeting the need for risk control. The platform can also avoid the drawbacks of long experimental cycles, empiricism in concentration selection, information omission and waste in traditional development methods for culture media formulations.
iBioG-Med contains three sub-models, I-SCR, I-FIT, and I-GEN. Variable selection model (I-SCR) can effectively describe the key components of the medium and establish the relationship between the medium components and technical objectives (e.g. antibody yield, antibody key quality attributes, etc.), which can be used to guide the subsequent optimization process and quality control of the medium production. Compared to traditional fitting models, the prediction model based on data distribution (I-FIT) has a higher expressivity and more accurate simulation and prediction of culture medium performance. The Learning model based on genetic algorithm (I-GEN) can be used to calculate formulation performance, greatly reducing workload and speeding up media development.
The iBioG-Med platform has successfully helped more than a dozen customers develop and optimize their culture media formulations. Some projects have been successfully marketed or entered clinical phases II and III. In addition, the iBioG-Med platform is used for the development of the DrivingM® standard catalog culture medium. iBioG-Med can iterate and upgrade the standard catalog media for each scenario to better meet the needs of our customers.
-
Quickly obtain the formula of culture media ( 3-6 months )
-
Identify key medium components and control range
-
Evaluate the stability of culture media formulations
- Variable selection model ( I-SCR )
-
It is highly accurate and has the ability to handle non-linear relationships.
It can indicate the direction of subsequent optimization.
It can identify key components that affect the medium product.
- Prediction model based on data distribution ( I-FIT )
-
It can effectively make predictions on experimental data with higher validity than traditional prediction models.
The error rate of the prediction model decreases gradually with data iteration, with an error rate of only 10-20%.
It can identify the range of key component concentrations (point → surface) and provide risk indication for media products.
It provides technical guarantees for self-learning model (I-GEN).
- Learning model based on genetic algorithm ( I-GEN )
-
≈300,000 formulations/day (I-GEN)
20-40 formulations for verification and validation
80-90% accessibility rate
It greatly reduces media development workload and accelerates development efficiency.
-
The platform has served 60+ projects in multiple fields
-
Formulation development and upgrade iteration of DrivingM ® standard catalog culture medium
-
Formulation development and upgrade iteration of Eden CD CHO Media
Let us contact you
We have a professional technical team to provide high quality one-stop cell culture service from process development and optimization, medium formulation design to medium processing and manufacturing for all biopharmaceutical companies.
If you have any question, please click "Quick Message" and leave your message, we will reply to you as soon as possible. If anything urgent, please call (86)21-68582660-2792.