Artificial Intelligence in Translational Medicine
Round 1
Reviewer 1 Report
The paper provides a literature review on the current state of AI proposals and applications in medicine. However, there are some issues to be resolved before it is ready for publication. The issues are the following.
- While Figure 2 provides the overview on the application categories that AI can be useful, the discussion in Section 2 does not appear to follow the overview described in Figure 2.
- It would be easier to read if there is a table to summarize (compare and contrast) the existing work in terms of dataset description, the objective of AI, AI algorithms and details, accuracy, etc. for each topic of discussion (section 2)
- Table 1 provides insightful information regarding the use of FDA-approved AI tools. However, as there is a long list of tools, it would be more readable to categorize them in some axis (e.g., organs, medical input devices, AI's input, etc.)
Overall, I think this is a very interesting paper that could help readers from the AI and/or the medical field to grasp the current state of the field but a revision is required to make the paper more readable.
Author Response
Referee 1
The paper provides a literature review on the current state of AI proposals and applications in medicine. However, there are some issues to be resolved before it is ready for publication. The issues are the following.
While Figure 2 provides the overview on the application categories that AI can be useful, the discussion in Section 2 does not appear to follow the overview described in Figure 2.
Authors: In section 2 are treated the main aspects of AI/ML that were also reported in the Figure 2. In fact, we discussed AI/ML in drug discovery, in diagnosis and prognosis, in clinical trials, in imaging, and in personalized medicine. For avoiding possible mistakes, we reported the following sentence before the figure The aspects illustrated in the Figure are discussed in Section 2. Furthermore, we prefer to maintain the Figure 2 in the introduction since it depicts the main aspects of AI/ML in medicine.
It would be easier to read if there is a table to summarize (compare and contrast) the existing work in terms of dataset description, the objective of AI, AI algorithms and details, accuracy, etc. for each topic of discussion (section 2)
Authors: we thank the referee for the suggestion, allowing us to improve the manuscript in terms of readability. We introduced 8 Tables at the end of each topic in the section 2 summarizing the AI approaches, targets, datasets, outcomes.
Table 1 provides insightful information regarding the use of FDA-approved AI tools. However, as there is a long list of tools, it would be more readable to categorize them in some axis (e.g., organs, medical input devices, AI's input, etc.)
Authors: we thank the referee for the comments. According to the request, we modified the Table 1 grouping the FDA-approved AI tools with respect to the medical field of application, and by alphabetical order.
Overall, I think this is a very interesting paper that could help readers from the AI and/or the medical field to grasp the current state of the field but a revision is required to make the paper more readable.
Authors: The authors thank the reviewer for the positive evaluation of the manuscript. All the corrections based on the reviewer’s comments have been done. Thanks to the referee suggestions, we were able to improve the quality of the manuscript, providing a more readable version of the manuscript. Accordingly, we have provided a tracked version of the manuscript for easily evaluating the changes done.
Reviewer 2 Report
The review of Dr. Brogi and Calderone concerns the use of artificial intelligence in translational medicine. The manuscript is well written and easy to read. It deals with all the aspects of AI and gives a comprehensible definition of all the types of machine and deep learning approaches.
The references cover a wide range of manuscripts
It deserves to be published unless one major issue:
- at lines 177-180 the authors stated: "Beyond the classical computational approaches in drug discovery, such as ligand- (mainly QSAR methods and pharmacophore modeling) [28-31] and structure-based strategies (mainly based on molecular docking and molecular dynamics) [32-35] or a combination of them [36-39]..." The paragraph is correct in its content but I found that the references cited are only about the authors of the review. I am not against the use of self-citations when appropriate, but in such a paragraph the authors should cite the manuscript of other eminent scientists involved in this field of studies
- Probably due to the length of the manuscript, there are a lot of typos, Please check the final version accurately
Author Response
The review of Dr. Brogi and Calderone concerns the use of artificial intelligence in translational medicine. The manuscript is well written and easy to read. It deals with all the aspects of AI and gives a comprehensible definition of all the types of machine and deep learning approaches.
The references cover a wide range of manuscripts
Authors: The authors thank the reviewer for the positive evaluation of the manuscript. All the corrections based on the reviewer’s comments have been done. Accordingly, we have provided a tracked version of the manuscript for easily evaluating the changes done.
It deserves to be published unless one major issue:
at lines 177-180 the authors stated: "Beyond the classical computational approaches in drug discovery, such as ligand- (mainly QSAR methods and pharmacophore modeling) [28-31] and structure-based strategies (mainly based on molecular docking and molecular dynamics) [32-35] or a combination of them [36-39]..." The paragraph is correct in its content but I found that the references cited are only about the authors of the review. I am not against the use of self-citations when appropriate, but in such a paragraph the authors should cite the manuscript of other eminent scientists involved in this field of studies
Authors: we thank the referee for the observation. Accordingly, we inserted other references from eminent scientists in the field, regarding computer-based approaches for drug discovery.
10.1039/D0CS00098A QSAR without borders
10.3390/molecules25061375 A Review on Applications of Computational Methods in Drug Screening and Design
10.1186/s13321-021-00537-9 QPHAR: quantitative pharmacophore activity relationship: method and validation
10.3390/ijms20112783 A Structure-Based Drug Discovery Paradigm
10.3389/fchem.2020.00343 Structure-Based Virtual Screening: From Classical to Artificial Intelligence
10.1021/acs.jmedchem.0c02227 Decision Making in Structure-Based Drug Discovery: Visual Inspection of Docking Results
10.1021/acs.jmedchem.5b01684 Role of Molecular Dynamics and Related Methods in Drug Discovery
10.3389/fmolb.2021.673773 Editorial: Molecular Dynamics and Machine Learning in Drug Discovery
10.3390/molecules25204723 Merging Ligand-Based and Structure-Based Methods in Drug Discovery: An Overview of Combined Virtual Screening Approaches
10.3389/fphar.2018.01416 Editorial: Chemoinformatics Approaches to Structure- and Ligand-Based Drug Design
-Probably due to the length of the manuscript, there are a lot of typos, Please check the final version accurately
Authors: the final version was carefully checked in order to avoid any possible typo errors.
Round 2
Reviewer 1 Report
The authors have done a great job in addressing the comments.
Reviewer 2 Report
No further comments