Digital to Biological Translation: How the Algorithmic Data-Driven Design Reshapes Synthetic Biology
Round 1
Reviewer 1 Report
Comments and Suggestions for AuthorsDear Authors,
The title of the review suggests how data-driven algorithms can impact synthetic biology.
However, what follows is mainly a compendium of definitions, concepts, and potential applications. It is appropiater for syntehtic biologist no specialists in these algorithms
Later, Table 1 presents ML/DL Contributions Across the Synthetic Biology DBTL Cycle, but these contributions are mentioned only as general cases, without discussing specific advances or application cases that have already been carried out.
Throughout the manuscript, there is no mention or discussion of cases where these algorithms have actually been applied to synthetic biology. For example, in the following reference:
Integrating Deep Learning and Synthetic Biology: A Co-Design Approach for Enhancing Gene Expression via N-Terminal Coding Sequences
There is a concrete application case with results that is worth reviewing in terms of its impact, and whether it opens new paradigms of study.
The manuscript’s mention of biology and relevant case studies is limited or almost nonexistent.
I believe it is important to include such examples in more detail, so that as a review, the article can better reflect what has been achieved and highlight the areas where improvements could be made.
Author Response
We thank and appreciate the valuable comments by Reviewer 1.
Comment 1: [Table 1 presents ML/DL Contributions Across the Synthetic Biology DBTL Cycle, but these contributions are mentioned only as general cases, without discussing specific advances or application cases that have already been carried out.]
Response 1: [Thank you for the valuable suggestion. We have substantially revised Table 1: AI/ML/DL contributions across the synthetic biology DBTL Cycle. We added a highly detailed column titled “Representative Studies" which directly links each AI contribution within the DBTL cycle to published literature.]
Comment 2: [Throughout the manuscript, there is no mention or discussion of cases where these algorithms have actually been applied to synthetic biology.]
Response 2: [Thanks for the comment. By focusing the discussion around the specific case studies (section 2.7) at page 10-11, the manuscript's engagement with biological mechanisms is now much deeper and more relevant. The revised manuscript now focused on the biological mechanisms and genotype-phenotype relationships revealed by AI, rather than just the algorithms themselves. The analysis of AI models' findings now explicitly discusses the resulting biological outcomes, such as enhanced protein expression, predicted gene essentiality, and improved product yield.]
Comment 3: [There is a concrete application case with results that is worth reviewing in terms of its impact, and whether it opens new paradigms of study.]
Response 3: [We have reviewed the application case with results, this work extends beyond its numerical results; it opens several new paradigms of study for rational design. Firstly, it demonstrates the efficacy of few-shot learning in synthetic biology, offering a powerful blueprint for overcoming the inherent data sparsity challenges of the field. Secondly, the model's ability to find optimal sequences suggests that DL is capturing a complex translational regulatory grammar that lies beyond traditional, easily interpretable biophysical models. This elevates AI from a mere predictive tool to an active co-designer, paving the way for autonomous systems that optimize entire regulatory cassettes.]
Comment 4: [The manuscript’s mention of biology and relevant case studies is limited or almost
nonexistent.]
Response 4: [We thank reviewer for this comment. We now introduce discussing high-impact case studies (section 2.7) at page 10-11 and then derive the broader implications and challenges from those examples. The manuscript now serves as an evidence-based critical review of the field's achievements, making it highly valuable to both synthetic biologists and computational specialists.]
Reviewer 2 Report
Comments and Suggestions for AuthorsDigital to Biological Translation: How the Algorithmic Data-Driven Design Reshapes Synthetic Biology
This review article examines the core tenets of AI, ML, and DL and its integration in synthetic biology workflows.
Reading from the abstract, it looks very broad and ambitious making it hard for the readers to focus. The strategic integration of AI, ML, and DL in synthetic biology workflow is already very broad. The additional topics on data sparsity, model interpretability, the “black box” problem, computational resource demands, and ethical considerations make the manuscript very broad.
It looks like the manuscript has several spacing issues. This, however, can be resolved easily during the copyediting stage of the manuscript.
Section 2 on Foundations of Artificial Intelligence, Machine Learning, and Deep 125 Learning is very long. I think the focus could just really be directly on the review of interest. Overall, I don’t think section 2 is necessary.
Please ensure to use acronyms consistently throughout the paper once already introduced.
I really think having a figure on the applications of synthetic biology in different fields would be great.
I would encourage the use of a graphical abstract in this paper.
Make sure to define all acronyms first prior to use and use that acronym all throughout the paper.
Table formatting is incorrect, I think.
There should be a figure showing how AI, ML, DL are integrated into synthetic biology workflow. The manuscript is very overwhelming for me.
I think the readers would be interested to know on how the integration process of AI, ML, and DL is done in the synthetic biology workflow. So far, many of the discussions are primarily based on the applications.
This review is very long.
There should be studies in table showing the integration of AI, ML, and DL in synthetic biology.
Author Response
We thank and appreciate the valuable comments by Reviewer 2. We have implemented targeted revisions to streamline the content and enhance its visual communication:
Comment 1: Reading from the abstract, it looks very broad and ambitious making it hard for the readers to focus.
Response 1: We have revised the abstract to clearly communicate the readers focus
Comment 2: Section 2 on Foundations of Artificial Intelligence, Machine Learning, and Deep 125 Learning is very long. I think the focus could just really be directly on the review of interest. Overall, I don’t think section 2 is necessary.
Response 2: The lengthy Section 2 has been deleted entirely.
Comment 3: There should be studies in table showing the integration of AI, ML, and DL in synthetic biology.
Response 3: Studies showing the integration of AI/ML/DL in synthetic biology in table 1. We introduced a detailed new section, “2.6. Integration Process of AI/ML/DL in the Synthetic Biology Workflow” which directly addresses the reviewer's query about how AI is integrated. This section explains the mechanism of integration at each DBTL stage.
Comment 4: I would encourage the use of a graphical abstract in this paper.
Response 4: Graphical abstract in added in the paper.
Comment 5: There should be a figure showing how AI, ML, DL are integrated into synthetic biology workflow.
Response 5: To address the request for illustrative figure we have created and integrated the essential figure 2 (AI-Driven DBTL Cycle) which serves as the central figure illustrating figure showing how AI/ML/DL integrated into synthetic biology workflow.
Comment 6: I really think having a figure on the applications of synthetic biology in different fields would be great.
Comment 6: Figure 3 (Applications Landscape) on the applications of synthetic biology in different fields is added.
Comment 7: It looks like the manuscript has several spacing issues. This, however, can be resolved easily during the copyediting stage of the manuscript. Please ensure to use acronyms consistently throughout the paper once already introduced. Table formatting is incorrect, I think.]
Response 7: All identified spacing issues have been resolved. We conducted a meticulous review to ensure all acronyms are defined upon first use and used consistently thereafter. All table formatting has been corrected to meet journal standards.
Round 2
Reviewer 1 Report
Comments and Suggestions for Authorshere is a clear improvement in the manuscript, and the new section (2.6. Integration Process of AI/ML/DL in the Synthetic Biology Workflow) is appropriate to enrich the discussion. However, this section still contains no citations. I recommend the following references to guide you in identifying additional applications that can support the discussion in the new section.
Azrag, M. A. K., Kadir, T. A. A., Kabir, M. N., & Jaber, A. S. (2019). Large-scale kinetic parameters estimation of metabolic model of Escherichia coli. International Journal of Machine Learning and Computing, 9(2), 160-167.
Cuperlovic-Culf, M. (2018). Machine learning methods for analysis of metabolic data and metabolic pathway modeling. Metabolites, 8(1), 4.
Cuperlovic-Culf, M., Nguyen-Tran, T., & Bennett, S. A. (2022). Machine learning and hybrid methods for metabolic pathway modeling. Computational Biology and Machine Learning for Metabolic Engineering and Synthetic Biology, 417-439.
Sieow, B. F. L., De Sotto, R., Seet, Z. R. D., Hwang, I. Y., & Chang, M. W. (2022). Synthetic biology meets machine learning. In Computational Biology and Machine Learning for Metabolic Engineering and Synthetic Biology (pp. 21-39). New York, NY: Springer US.
Li, Y., Wu, F. X., & Ngom, A. (2018). A review on machine learning principles for multi-view biological data integration. Briefings in bioinformatics, 19(2), 325-340.
Author Response
Comment 1: [here is a clear improvement in the manuscript, and the new section (2.6. Integration Process of AI/ML/DL in the Synthetic Biology Workflow) is appropriate to enrich the discussion. However, this section still contains no citations. I recommend the following references to guide you in identifying additional applications that can support the discussion in the new section.]
Response 1: [Thanks for the comment. The references has been added in the new section (2.6.)]
