This study combined the transcriptome of human chronic kidney disease (CKD) with two machine learning algorithms to screen for high- and low-risk genes associated with CKD, and to construct a risk prediction model. The main research method involved using machine learning in transcriptomics to build a predictive model. Has similar research been reported before? How effective is its predictive performance?
In this article, four different gene set scoring methods were used. Is this necessary?
In Figure 4A, the highest predictive accuracy for CKD progression was 0.687. Can this be further improved?
