Next Article in Journal
Characterization of Salivary Microbiota in Japanese Patients with Oral Cancer
Next Article in Special Issue
Preconditioning with Rapamycin Improves Therapeutic Potential of Placenta-Derived Mesenchymal Stem Cells in Mouse Model of Hematopoietic Acute Radiation Syndrome
Previous Article in Journal
Neutrophil Extracellular Traps in Atherosclerosis: Research Progress
Previous Article in Special Issue
Transcriptome-Wide Insights: Neonatal Lactose Intolerance Promotes Telomere Damage, Senescence, and Cardiomyopathy in Adult Rat Heart
 
 
Article
Peer-Review Record

Comparative Study of Deep Transfer Learning Models for Semantic Segmentation of Human Mesenchymal Stem Cell Micrographs

Int. J. Mol. Sci. 2025, 26(5), 2338; https://doi.org/10.3390/ijms26052338
by Maksim Solopov 1, Elizaveta Chechekhina 2, Anna Kavelina 1, Gulnara Akopian 1, Viktor Turchin 1, Andrey Popandopulo 1,*, Dmitry Filimonov 1 and Roman Ishchenko 1
Reviewer 1: Anonymous
Reviewer 2: Anonymous
Int. J. Mol. Sci. 2025, 26(5), 2338; https://doi.org/10.3390/ijms26052338
Submission received: 28 January 2025 / Revised: 26 February 2025 / Accepted: 4 March 2025 / Published: 6 March 2025

Round 1

Reviewer 1 Report

Comments and Suggestions for Authors

Dear Authors,
the analysis of microscopic images of live cell cultures on ideal cell culture surfaces, especially for stem cells and primary, differentiated cells, is highly dependent on many factors. Even with standardised laboratory protocols in the form of SOPs and under GLP conditions, it also depends on the experience of the respective personnel. The time factor for the evaluation is also considerable. Today, it is imperative to implement modern automation and AI applications.
Your comparative study of ‘deep transfer learning models for semantic segmentation’, here of hMSCs, is enormously beneficial for the global life sciences community for an international harmonisation of image data evaluations and their standardisation.
This work is very well and clearly designed, explained in a largely plausible manner, and the results are well suited for applications to more complex, morphologically inhomogeneous cell types. The authors' conclusion regarding the usability of the U-net model is consistent with our own experiences.
There are a few notes from me for improvement, which I have added as comments to the uploaded version.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Comments and Suggestions for Authors
  • The authors need to provide details on the source and type of the human MSC used in this study. The authors used 20% of FBS for MSC culture, which is a high percentage. Any reason?
  • The authors applied four neural network models for image segmentation—U-Net, DeepLabV3+, Mask R- 265 CNN and SegNet. The rationale of selecting those platforms needs to be clearly explained. Is there any gold standard method to be used as a control?
  • The authors need to provide cell pictures and conventional methods for analyzing cell phenotype in comparison to their platforms.
Comments on the Quality of English Language

Need to improve the quality of English and check the grammatical mistakes and typos.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Back to TopTop