融合人工智能和自动化的蛋白质设计表征实验教学实践

[Objective] With rapid developments in synthetic biology, guided by engineering principles, there is a growing emphasis on the top-down development of biotechnology to address major societal challenges. Simultaneously, the swift progress in computer science has promoted the effective development of varied disciplines in many fields. The seamless integration of artificial intelligence (AI) and automation in biochemical engineering necessitates educational approaches that prepare students for this dynamic and ever-evolving landscape. Thus, we developed “Protein computational design and characterization,” a laboratory course for biochemical engineering students in the first year of their graduation that combines AI-driven computational protein design with hands-on experimental techniques. [Methods] Students enrolling typically possess a strong theoretical foundation in mathematics and chemistry but lack experience in automatic instrument operation and programming, with many not being preexposed to biochemical experiments. Hence, this course integrates AI-based computational protein design with wet-lab techniques such as cell-free protein synthesis and high-throughput screening by employing advanced instruments like automatic liquid handling workstations. A software application with a user-friendly graphic interface was developed to assist students in navigating protein design workflow, including motif identification and selection, backbone and sequence design, and scoring and analysis. This workflow integrates AI-based protein design programs such as RFdiffusion and ProteinMPNN, protein structure prediction tools such as OmegaFold, and protein visualization tools. The high-throughput screening platforms, including an automatic liquid handling workstation and a plate reader, enable the parallel screening of several predesigned proteins and perform activity assays employing multiple substrates or protein concentrations. Additionally, the cell-free protein synthesis technique allows for in vitro protein synthesis and purification within a single workday starting from synthetic DNA, thereby bypassing traditional transformation and cell culture steps. Additionally, green fluorescent protein is utilized as a protein label during gene design to facilitate the fluorescence-based identification of protein expression and purification during screening. Students would build model proteins with a small molecular weight and simple structure to simulate the function of classical proteins, such as RNase A, through iterative experiments and predictive calculations. Across four course units, the students will be systematically and progressively exposed to AI, fundamental laws of biology, high-throughput techniques for experimental automation, and protein experimental methodologies, enabling an understanding of these. [Results] By combining computational (“dry lab”) and experimental (“wet lab”) methods, the course serves as a comprehensive platform where students can systematically learn fundamental biochemical operations while fostering interdisciplinary thinking. Students are exposed to the entire protein engineering workflow—from computer design and gene synthesis to protein expression, purification, and functional characterization. The inclusion of large automated instruments not only enhances their hands-on experience with advanced equipment but also prepares them for future research in intelligent and automated biochemical laboratories. [Conclusions] The course “Protein computational design and characterization” effectively develops the experimental skills, problem-solving ability, and resilience of students helping them in scientific research. It serves as an effective model for integrating cutting-edge technologies and interdisciplinary approaches applicable to graduate education in biochemical engineering. This paper provides an overview of the course design, including its conceptual framework, experimental procedures, teaching methods, and assessment strategies. We discuss the challenges encountered and the results achieved, highlighting how integrating intelligent automation into biochemical engineering education can enhance the learning experience of students and better prepare them for their careers.