Computational modeling and machine learning for high content screening of 3D cell culture environments for biopharmaceutical discovery

Project description

Microfluidic based high content screening of cells embedded in 3D hydrogels provides a powerful alternative to traditional suspension or 2D culture methods, offering higher quality cell populations at lower cost. Our collaborative team has engineered high throughput microchip platforms where single cells can be live imaged to track stem cell expansion in 3D hydrogels, creating the first 3D hematopoietic stem cell (HSC) platforms. This project seeks to advance that foundation by building a next generation high throughput screening pipeline that integrates microfluidic live cell imaging and experimental data with computational modeling and machine learning. By constructing mathematical representations of the underlying biological processes, we can simulate and predict culture dynamics, quantify key variables, and map the interactions between process indicators and control parameters. Such insights will support more rational process design and optimization, improve control strategies, and reduce the need for repetitive experimental trials.

Aims: 1. Develop mechanistic, machine learning based, and hybrid models to investigate cell proliferation within hydrogels, coupled with live cell imaging data (image processing). 2. Employ process systems engineering approaches, including surrogate modeling, process monitoring, optimization, and control, to enhance cell production, improving efficiency while reducing costs. 3. Experimentally validate computational predictions using custom designed microfluidic chip platforms.

Methodology: 1. At IIT Delhi, computational frameworks will be established to capture the key factors influencing cell growth and to predict optimal culture strategies. Advanced ML methods, guided by biological insight, will be explored alongside mechanistic models. A process monitoring and optimization framework will be developed to track and refine critical process variables. 2. At UQ, the student will receive experimental training and perform data collection using microfluidic chip systems. The computational predictions will be tested and validated experimentally in close collaboration with UQ researchers.

Outcomes

* Comprehensive literature review of mechanistic, machine learning, and hybrid modeling approaches for cell culture processes. * Development of computational models to describe cell proliferation and culture dynamics. * Application of data preprocessing and image analysis techniques for extracting relevant biological features. * Creation of surrogate models and optimization frameworks to improve process efficiency and reduce experimental effort. * Integration of process monitoring and/or control strategies for predictive cell culture management. * Experimental validation of computational predictions using custom microfluidic platforms at UQ. * Development of a user-friendly graphical interface to integrate modeling, monitoring, and optimization tools. * Publications in high-impact journals and a doctoral thesis consolidating the research contributions.

Essential capabilities

Background in chemical engineering, biotechnology, biomedical engineering, mathematics, physics, computer science, or related disciplines. Solid foundation in computer programming. Willingness to learn new skills across both computational and experimental domains.

Desireable capabilities

Experience in computational modeling, machine learning, image analysis (e.g., in Python). Familiarity with cell culture or microfluidics.

Expected qualifications (Course/Degrees etc.)

At least a Bachelor’s degree in an aligned field such as chemical engineering, biotechnology, biomedical engineering, mechanical engineering, electrical engineering, mathematics, physics, computer science, or related disciplines.