Foreword

From the release chatGPT on November 30, 2022 to the release of Wen Xin, from the Co-pilot launched by Microsoft to the Firefly launched by Adobe, we seem to be at the forefront of this wave of AI technology. The big language model represented by ChatGPT has formed a new paradigm of artificial intelligence, and has also boosted the development of different academic fields. The release of Alphafold2 in 2021 and the open source of code on Github became a milestone event in the field of life sciences that year. At the beginning of this year, the article on the protein language model ProGen published on Nature Biotechnology and the article on the generative artificial intelligence (AI) model Chroma published on Nature in November that can design proteins not found in nature make us deeply feel the infinite possibilities of AI technology in synthetic biology. With the rapid development of artificial intelligence (AI) technology and cell-free protein expression (CFPS) technology, their applications in protein prediction and manufacturing are more closely and widely combined.

1. AI and CFPS Technology Development

Research in the field of AI is becoming more and more popular. With the gradual improvement of hardware equipment and the optimization of algorithms (including machine learning, deep learning, natural language processing, and computer vision), it has become an indispensable part of human production and life. Among them, deep learning language models have shown prospects in various biotechnology applications, including protein design and engineering. Prior to the AI revolution, protein design methods were limited to generating designs based on existing proteins in nature, which may have sampled only a small part of the protein landscape in nature. By contrast, AI approaches extend the range of desired functions and properties beyond what is already being realized in nature by designing proteins from scratch.

In recent years, CFPS technology has developed rapidly and has become a platform technology. It mimics whole cell transcription and translation processes in an in vitro controlled environment without the need for an intact living cell, and allows detailed in-depth studies of individual components and reaction networks. Compared with the traditional in vivo protein expression system, CFPS system has many advantages, such as controllable reaction conditions, fast synthesis, high protein yield, no cell toxicity and so on. It has been successfully used in the synthesis of drug proteins and membrane proteins and other difficult to express proteins, and has broad prospects in basic science research and applied science research.

Joint application of 2. AI and CFPS

The combination of cell-free protein expression and AI will promote the rapid development of synthetic biology, drug development, protein engineering and many other frontier fields, which has broad application prospects and commercial value. Amir Pandi et al. built an LSTM model (called LSTM_Pep) to generate antimicrobial peptides de novo and fine-tuned the model to generate peptides with specific expected therapeutic benefits. In addition, the researchers developed a deep learning-based protein-peptide prediction model (DeepPep) for rapid screening of peptides generated by a given target. The design-build-test cycle in AMP development is greatly advanced by the use of CFPS. The study reveals a technology platform for developing deep learning-based methods combined with CFPS, showing how artificial intelligence can help design synthetic antimicrobial peptides (AMPs) from scratch. In addition, it also provides us with a revelation that CFPS can fully realize the potential of machine learning in protein design, especially in the context of the declining cost of DNA synthesis. The combination of deep learning and CFPS provides a fast, low-cost and efficient method for peptide production and screening. This will hopefully increase the discovery and development of peptide-based drug candidates in the future.