Next Article in Journal
Identification of a Selective Inhibitor of Human NFS1, a Cysteine Desulfurase Involved in Fe-S Cluster Assembly, via Structure-Based Virtual Screening
Next Article in Special Issue
Machine Learning Identification of Neutrophil Extracellular Trap-Related Genes as Potential Biomarkers and Therapeutic Targets for Bronchopulmonary Dysplasia
Previous Article in Journal
Increased Plasma Levels of Thrombin-Cleaved Osteopontin in Patients with Delayed Cerebral Infarction After Aneurysmal Subarachnoid Hemorrhage
Previous Article in Special Issue
Exploring Potential Medications for Alzheimer’s Disease with Psychosis by Integrating Drug Target Information into Deep Learning Models: A Data-Driven Approach
 
 
Article
Peer-Review Record

SST-ResNet: A Sequence and Structure Information Integration Model for Protein Property Prediction

Int. J. Mol. Sci. 2025, 26(6), 2783; https://doi.org/10.3390/ijms26062783
by Guowei Zhou 1,2,†, Yanpeng Zhao 2,3,†, Song He 2,* and Xiaochen Bo 1,2,*
Reviewer 1:
Reviewer 2: Anonymous
Int. J. Mol. Sci. 2025, 26(6), 2783; https://doi.org/10.3390/ijms26062783
Submission received: 24 February 2025 / Revised: 14 March 2025 / Accepted: 18 March 2025 / Published: 19 March 2025

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

The manuscript reports the development of a novel end-to-end deep learning framework for protein property prediction, integrating information from both protein sequences and 3D structures. This is based on recently reported pre-trained multimodal language models such as ProSST.

The modern state-of-the-art is well introduced. In the results, the proposed framework, called SST-ResNet, is communicated in a clear manner using the schema and a detailed description of each step of the prediction process. The evaluation metrics (scores) are well explained and the performance framework is evaluated for each method used and the ablation experiments were conducted.

In addition, the ability of SST-ResNet to learn latent relationships between protein sequences and structures was explored on adenosine triphosphate (ATP) and ribosomal RNA (rRNA). The predicted results match the empirical structures of the sample proteins that are either able to bind rRNA or not.

The methods used are well described for each step of the proposed framework.

 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors

This manuscript is well-organized and demonstrates a novel protein property prediction framework, SST-ResNet. This method utilizes protein structural information in addition to protein sequences and outperforms existing platforms. The methods, applications, and prediction results are well-presented and explained. I hope the web server will be available soon.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Back to TopTop