**5. Conclusions**

The machine learning algorithms show good abilities to differentiate various groups across large data sets. Full proteomics data obtained by mass spectrometry can be very useful for this approach to medical diagnosis. These data contain general information about changes in normal body processes in those cases when common diagnostic approaches using single biomarkers or even panels of biomarkers do not work well enough. With the testing machine learning models on proteomics data obtained from the plasma and urine of patients of three types of CKD, the best results were obtained using the nearest-neighbor algorithm. In this case, according to the proteomics data of plasma, the two groups of patients with diabetic nephropathy and glomerulonephritis are well separated from the group of healthy people. On the other hand, a less presented group of patients with hypertensive nephropathy is better isolated from groups of patients of the two other CKD types by the "one against all" method based on the urine proteome data set.

The further development of the approach presented here may help to avoid invasive intervention and contraindicated intervention in some cases for the verification of the glomerulonephritis subtypes, which is currently performed only by kidney biopsy and microscopic morphological confirmation. The diagnosis of hypertensive and diabetic nephropathy at an early stage also remains relevant and the capabilities of machine learning methods based on proteomics data may be useful for this purpose.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/1422-0067/21/13/4802/s1, Table S1. Plasma and urine proteins LFQ values.

**Author Contributions:** Conceptualization, methodology, mass spectrometry, software calculations, writing—original draft preparation, Y.E.G. and D.V.V.; blood and urine samples collection and preparation, medical investigations, I.A.L., M.L.R., S.A.V., and S.L.G.; medical investigations, writing—review and editing, T.N.Z. and M.M.P.; mass spectrometry experiments, Z.M.; supervision, review, and editing of the final version of the manuscript, M.V.B. and A.S.K. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** The authors sincerely thank Ivan Denisov for help in preparing the manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.
