Computational Molecular Biology 2025, Vol.15, No.3, 112-121 http://bioscipublisher.com/index.php/cmb 119 key points for verification, which is much more efficient and may also uncover regulatory factors that are difficult to discover through traditional experiments. It can be said that sequence function prediction driven by deep learning is tightening the connection between genotypes and phenotypes, bringing more precise and efficient research methods to medicine and biology. If we string together all the previous discussions, the potential of deep learning to predict gene expression using genomic sequences is already quite obvious and is still moving forward. From data preparation, model design to result interpretation, this method is becoming increasingly mature. It can capture sequence features that are difficult to identify by traditional methods, demonstrating unprecedented precision in the study of gene expression regulation and bringing new ideas to functional genomics. The application prospects are also quite broad: from the screening of disease risk variations to molecular breeding of crops, it can be put to good use. However, for such technologies to play a greater role, joint efforts from both the academic and industrial sectors are still needed. In research, it is necessary to continuously accumulate multi-dimensional high-quality data and also develop more efficient and easier-to-understand model algorithms. In terms of policy, efforts should be made to promote the cultivation of interdisciplinary talents, better integrate life sciences and artificial intelligence, establish an open and shared genome and expression database, and at the same time formulate relevant AI application norms to ensure the reliability and long-term usability of prediction results. As long as both technology and policy keep pace, deep learning-driven gene expression prediction is expected to truly transform the landscape of life science research and take precision medicine and biotechnology innovation to a new level. Acknowledgments I would like to express my heartfelt thanks to all the teachers who have provided guidance for this study. Conflict of Interest Disclosure The author affirms that this research was conducted without any commercial or financial relationships that could be construed as a potential conflict of interest. References Almotairi S., Badr E., Abdelbaky I., Elhakeem M., and Abdul Salam M., 2024, Hybrid transformer-CNN model for accurate prediction of peptide hemolytic potential, Scientific Reports, 14(1): 14263. https://doi.org/10.1038/s41598-024-63446-5 Avsec Ž., Agarwal V., Visentin D., Ledsam J., Grabska-Barwinska A., Taylor K., Assael Y., Jumper J., Kohli P., and Kelley D., 2021, Effective gene expression prediction from sequence by integrating long-range interactions, Nature Methods, 18(10): 1196-1203. https://doi.org/10.1038/s41592-021-01252-x Beer M., and Tavazoie S., 2004, Predicting gene expression from sequence, Cell, 117(2): 185-198. https://doi.org/10.1016/S0092-8674(04)00304-6 Chen R., Dai R., and Wang M., 2020, Transcription factor bound regions prediction: Word2Vec technique with convolutional neural network, Journal of Intelligent Learning Systems and Applications, 12(1): 1-13. https://doi.org/10.4236/jilsa.2020.121001 Chen Y., Li Y., Narayan R., Subramanian A., and Xie X., 2016, Gene expression inference with deep learning, Bioinformatics, 32(12): 1832-1839. https://doi.org/10.1101/034421 Choong A., and Lee N., 2017, Evaluation of convolutional neural networks modeling of DNA sequences using ordinal versus one-hot encoding method, In: 2017 International Conference on Computer and Drone Applications (IConDA), IEEE, pp.60-65. https://doi.org/10.1101/186965 Dong G., Wu Y., Huang L., Li F., and Zhou F., 2024a, TExCNN: Leveraging pre-trained models to predict gene expression from genomic sequences, Genes, 15(12): 1593. https://doi.org/10.3390/genes15121593 Dong W., Zhang J., Dai L., Chen J., Wu H., He R., Pang Y., Wang Z., Jian F., Ren J., Liu Y., Tian Y., Liu S., Zhao X., and Xie X., 2024b, Mapping eukaryotic chromatin accessibility and histone modifications with DNA deaminase, bioRxiv, 24: 630236. https://doi.org/10.1101/2024.12.24.630236 Dorka N., Welschehold T., and Burgard W., 2023, Dynamic update-to-data ratio: minimizing world model overfitting, arXiv, 2303: 10144. https://doi.org/10.48550/arXiv.2303.10144 Drusinsky S., Whalen S., and Pollard K., 2024, Deep-learning prediction of gene expression from personal genomes, bioRxiv, 27: 605449. https://doi.org/10.1101/2024.07.27.605449
RkJQdWJsaXNoZXIy MjQ4ODYzNA==