Proteomics-Based Machine Learning Approach as an Alternative to Conventional Biomarkers for Differential Diagnosis of Chronic Kidney Diseases

Round 1

Reviewer 1 Report

The authors describe  a Proteomics-Based Machine Learning Approach as an Alternative to Conventional Biomarkers for Differential Diagnosis of Chronic Kidney Diseases. It is an interesting manuscript, which could be improved after revision.

 Major comments:

Abstract: provides a good summary of the article

Introduction: Generally good, but with respect to the opinions of the  
authors regarding biopsy 55-58 , this is not supported by the majority of clinical data

The introduction is clear, useful in setting the scene and  understandable but there are  few, very minor, errors in sentence construction

Materials and methods section:

Please provide more data on patients groups

1. Age, stage of renal disease, vintage of CKD (in advanced CKD, renal dysfunction and reduced clearance may affect proteomics results)
2. Proteinuria or not -level of proteinuria
3. In different glomerulopathies, on active disease stage, selective proteinuria may be noticed, how this may influence blood/ urine proteomics?
4. Patients with glomerulopathies/diabetes were not also hypertensive?
5. Hypertensive group (cause of hypertension/ controlled or uncontrolled hypertension)
6. Urine of controls should be also analyzed

Results section: it would be interesting to add more detailed data on differences on proteome profile between different patients groups

Conclusions: Whilst I understand the desire  to try to derive meaning,  the authors would be better avoiding not particularly useful, conclusions regarding renal biopsy. Renal biopsy apart from diagnosis provides also important information regarding the severity of renal involvement. Therefore I believe that renal biopsy cannot avoided in many cases.

Minor comments:

The manuscript will require only  minor copy editing during production to formalise the English.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

Reviewing:

The authors present an interesting solution to a relevant question in clinical practice. In some cases, the distinction between diabetic/vascular nephropathy and primary glomerular disease may require a kidney biopsy. A noninvasive tool could be helpful to avoid biopsy in some cases.

Minor remarks:

-Line 104, 240 the term of hypertonic should be replaced by hypertensive.

-The authors should mitigate the sentence concerning the complications of kidney biopsy 

Major remarks:

How the diagnosis of diabetic nephropathy, hypertensive nephropathy, or glomerulonephritis has been made? Has a renal biopsy been performed for each patient? What was the estimated glomerular filtration rate of these patients? In the G group what type of glomerulonephritis were included?

A Table summarizing the principal characteristics of each group should be added for more clarity

The diagnosis hypertensive nephropathy (nephroangiosclerosis) is clinical and proteinuria is absent in most of the case. The interest of this kind of tools seems to be limited in this population. 

The distinction between primary glomerulonephritis and diabetic nephropathy is more relevant, as is based on histological findings in some cases. However, the majority of the patients presenting diabetic nephropathy (notably type 1) don’t need a kidney biopsy.  

The major limitation of this study is that diabetic patients probably present particular proteomic phenotypes distinguish them from other population of patients. What is the impact of diabetes on the distinction between group D and G? A study comparing 2 groups of diabetic patients, one presenting diabetic nephropathy and a second presenting primary glomerulonephritis, would be probably more informative. This limit should be discussed extensively.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The major and minor points of my review have been sufficiently  responded

I have no further comments

Reviewer 2 Report

The major limitation is now discussed. I have no further comment.