Our lives have gradually become inseparable from AI. DUT-China has innovatively summarized all the scenarios where AI is used in the process of our project advancement. We hope to reflect on the pros and cons of AI usage through this, and provide a reference for the iGEM community's future use of AI.
AI-based protein language model for mutation effect prediction: Using AI protein language models such as SaprotHub for zero-shot mutational effect prediction, one can input amino acid sequences with structural information and perform neural network evaluation for each amino acid residue, obtaining an average output score. This method can quantify the impact of single or multiple amino acid changes on the entire protein. For instance, scoring and analyzing the full-site mutations of the Mad7 protein can assist scientists in quantifying the impact of individual mutation sites and serve as a predictive basis for wet experiments.
AI can assist researchers in efficiently designing experimental ideas, reducing the time spent on searching and reading a large amount of data. For instance, when faced with the problem of optimizing the binding ability of proteins and small molecules, AI can quickly decompose the problem, select targets and ligands, conduct molecular docking simulations to predict affinity and binding conformation, use kinetic models to observe behavior and stability under physiological conditions, calculate free energy, and thus iteratively optimize and conduct experimental verification.
AI can assist in writing code, efficiently processing and analyzing data, and creating charts. For instance, Python code written with AI can merge multiple data tables, match and analyze corresponding data from different tables, and generate data charts, thereby enhancing the efficiency and accuracy of data processing.
Using AI to assist in rapid literature reading and comprehension: AI tools can help quickly locate relevant literature and interpret and extract its content. For instance, when studying the toxicity of SacB, AI can search for literature on related mechanisms, swiftly distill key content and necessary information, and assist researchers in better understanding and summarizing the research topic.
AI can assist in translating and optimizing literature in different languages, transforming colloquial expressions into more academic expressions, and integrating and constructing highly readable text frameworks. This is particularly useful for researchers when writing scientific papers or reports, as it can save a significant amount of time in writing and proofreading.
AI predicts laboratory risks: AI tools can assist in analyzing unknown risks in experiments, helping teams comprehensively consider safety precautions. We have referenced AI's safety analysis of CRISPR-Cas gene editing in the Safety Form.
Although AI can bring us the aforementioned conveniences, we cannot fully rely on AI for iGEM activities. The immaturity of AI technology can easily lead to insufficient data accuracy. The abuse of AI may trigger social exclusion and may also cause ethical and legal issues. We must dialectically view the development of AI and use it in a reasonable manner.