The Frontier Exploration of Algorithm Innovation and Experimental Verification in Intelligent Protein Design

Yi Sui

doi:10.54691/jmxh2s40

Authors

Yi Sui

DOI:

https://doi.org/10.54691/jmxh2s40

Keywords:

Protein Design; Algorithm Innovation; Experimental Verification; DeepThermoNet; Deep Learning.

Abstract

Intelligent protein design is a frontier topic in the cross field of modern biotechnology and AI. Through the combination of algorithm innovation and experimental verification, it breaks through the limitations of traditional protein design. In this paper, the progress of algorithm innovation in intelligent protein design is summarized, especially the application of advanced algorithms such as deep learning, generative model and reinforcement learning in protein structure prediction, function optimization and interaction analysis. Taking DeepThermoNet, a deep learning algorithm, as an example, the effect of protein mutant designed by DeepThermonet in improving the thermal stability of β -glucosidase was verified by experiments. The results showed that the mutant designed by the algorithm group was significantly better than the mutant designed by the traditional method in melting temperature (Tm) and enzyme activity retention rate. The experimental verification not only proves the effectiveness of the algorithm design, but also optimizes the algorithm model through feedback, forming a closed loop of "algorithm design-experimental verification-model optimization". This paper further discusses the interactive relationship between algorithm innovation and experimental verification, looks forward to the future development direction of intelligent protein design, including interdisciplinary integration, new algorithm development and data resource expansion, and points out the limitations of current research and the key direction of future work. Intelligent protein design is expected to provide new theoretical and technical support for drug research and development, biocatalyst development and biomaterial design, and promote innovation and development in related fields.

Downloads

Download data is not yet available.

References

[1] Crisman, E. , Duarte, P. , Dauden, E. , Cuadrado, A. , María Isabel.Rodríguez-Franco, & Manuela G.López, et al. (2023). Keap1-nrf2 protein-protein interaction inhibitors: design, pharmacological properties and therapeutic potential. Medicinal research reviews, 43(1), 237-287.

[2] Wu, Z. , Johnston, K. E. , Arnold, F. H. , & Yang, K. K. (2021). Protein sequence design with deep generative models. Current Opinion in Chemical Biology, 65(15), 18-27.

[3] Jií.Zahradník, & Schreiber, G. (2021). Protein engineering in the design of protein-protein interactions: sars-cov-2 inhibitors as a test case. Biochemistry, 60(46), 3429-3435.

[4] Courbet, A. , Hansen, J. , Hsia, Y. , Bethel, N. , Park, Y. J. , & Xu, C. , et al. (2022). Computational design of mechanically coupled axle-rotor protein assemblies. Science (New York, N.Y.), 376(6591), 383-390.

[5] Koroleva, E. V. , Ermolinskaya, A. L. , Ignatovich, Z. V. , Kornoushenko, Y. V. , Panibrat, A. V. , & Potkin, V. I. , et al. (2024). Design,in silicoevaluation, and determination of antitumor activity of potential inhibitors against protein kinases: application to bcr-abl tyrosine kinase. Biochemistry (Moscow), 89(6), 1094-1108.

[6] Kortemme, T. (2024). De novo protein design—from new structures to programmable functions. Cell, 187(3), 526-544.

[7] Reardon, S. (2024). Five protein-design questions that still challenge ai. Nature, 635(8037), 246-248.

[8] Chu, A. E. , Lu, T. , & Huang, P. S. (2024). Sparks of function by de novo protein design. Nature biotechnology, 42(2), 203-215.

[9] Doerr, A. (2024). Protein design: the experts speak. Nature Biotechnology, 42(2), 175-178.

[10] Chen, Z. , Ji, M. , Qian, J. , Zhang, Z. , Zhang, X. , & Gao, H. , et al. (2024). Probid-net: a deep learning model for protein–protein binding interface design. Chemical Science, 15( 47), 19977-19990.

[11] Wand, A. J. (2024). How to design a protein that can be switched on and off. Nature, 632(8026), 741-742.

[12] Chen, Z. (2023). Protein circuit design using de novo proteins. Trends in biotechnology, 41(5), 593-594.

[13] Shintaro, M. , Naohiro, K. , & Sugiki ToshihikoNagashima ToshioFujiwara ToshimichiTatsumi-Koga RieChikenji GeorgeKoga Nobuyasu. (2023). Exploration of novel αβ-protein folds through de novo design. Nature structural & molecular biology, 30(8), 1132-1140.

[14] Huang, B. , Xu, Y. , Hu, X. , Liu, Y. , Liao, S. , & Zhang, J. , et al. (2022). A backbone-centred energy function of neural networks for protein design. Nature, 602(7897), 523-528.

[15] Pulavarti, S. V. S. R. K. , Maguire, J. B. , Yuen, S. , Harrison, J. S. , Premkumar, L. , & Weiss, T. M. , et al. (2022). From protein design to the energy landscape of a cold unfolding protein. The Journal of Physical Chemistry B, 126(6), 1212-1231.

[16] Kucera, T. , Togninalli, M. , & Meng-Papaxanthos, L. (2022). Conditional generative modeling for de novo protein design with hierarchical functions. Bioinformatics (Oxford, England), 38(13), 3454-3461.

[17] Ferguson, A. L. , & Ranganathan, R. (2021). 100th anniversary of macromolecular science viewpoint: data-driven protein design. ACS Macro Letters, 10(3), 327-340.