Machine-learning-assisted nanopore sensing solution for the determination of matrix metalloproteinase

Diabetic kidney disease (DKD) is a serious complication of diabetes patients with long time duration, presenting with albuminuria and/or a reduced estimated glomerular filtration rate (eGFR), and without symptoms of other primary causes of kidney injury. Clinical studies showed matrix metalloproteinase 2 (MMP2) is the potential indicator for DKD diagnosis. However, the typical measurement of MMP2 is complicated and time-consuming. Therefore, it is necessary to develop an easy and reliable approach for MMP2 detection. Herein, we proposed a reliable and easy-to-use nanopore solution for the quantitative measurement of MMP2 at the single-molecule level using α-hemolysin nanopore. Assisted by machine learning, the peptide substrate and peptide products digested by MMP2 were classified with 100 % accuracy. The quantitative range of MMP2 concentration was 50–400 ng/ml. We further investigated the inhibitory effects of MMP2 activity by different chemicals including Cu²⁺, Ni²⁺, Zn²⁺, EDTA, and its inhibitor GM6001. Finally, MMP2 measurement was explored in the presence of simulated urine. Our research provides a new solution of quantification of MMP2 activity combined with machine learning for DKD diagnosis.