Despite tremendous developments in continuous blood glucose measurement (CBGM) sensors, they are still not accurate for all patients with diabetes. As glucose concentration in the blood is <1% of the total blood volume, it is challenging to accurately measure glucose levels in the interstitial fluid using CBGM sensors due to within-patient and between-patient variations. To address this issue, we developed a novel data-driven approach to accurately predict CBGM values using personalized calibration and machine learning. First, we scientifically divided measured blood glucose into smaller groups, namely, hypoglycemia (<80 mg/dL), nondiabetic (81–115 mg/dL), prediabetes (116–150 mg/dL), diabetes (151–181 mg/dL), severe diabetes (181–250 mg/dL), and critical diabetes (>250 mg/dL). Second, we separately trained each group using different machine learning models based on patients’ personalized parameters, such as physical activity, posture, heart rate, breath rate, skin temperature, and food intake. Lastly, we used multilayer perceptron (MLP) for the D1NAMO dataset (training to test ratio: 70:30) and grid search for hyperparameter optimization to predict accurate blood glucose concentrations. We successfully applied our proposed approach in nine patients with type 1 diabetes and observed that the mean absolute relative difference (MARD) decreased from 17.8% to 8.3%.

Background:: Multi drug-resistant tuberculosis is a major health threat to humans. Whole genome sequencing of several isoniazid (INH) resistant strains of M. tuberculosis revealed mutations in several genes. Rv1592c was demonstrated as lipolytic enzyme and its expression was up-regulated during isoniazid (INH) treatment. The valine at position 430 of Rv1592c was mutated to alanine frequently in the INH resistant strain of M. tuberculosis. Methods: In this report, an array of computational approaches was used to understand the role of Val430-Ala mutation in Rv1592c in INH resistance. The impact of mutations on structural stability and degree of INH modification was demonstrated using the molecular dynamics method. The mutation in the Rv1592c gene at V430 position was created by the PCR primer walking method. Mutant and wild type gene was cloned into E. coli-mycobacteria shuttle vector (pVV-16) and expressed in Mycobacterium smegmatis system. The isoniazid susceptibility assay was performed by agar plate culture spot and CFUs count assay. Results: This study demonstrated that the Val430 in Rv1592c makes the part of flap covering the substrate binding cavity. Mutation at Val430-Ala in Rv1592c caused the displacement of the flap region, resulting in uncovering a cavity, which allows accessibility of substrate to the active site cleft. The Val430-Ala mutation in Rv1592c created its structure energetically more stable. RMSD, RMSF and Rg simulation of mutant maintained overall stability throughout the simulation period while the native protein displayed comparatively more fluctuations. Moreover, docking studies showed that INH was bound into the active pocket of the mutant with considerable binding energy (−6.3 kcal/mol). In order to observe constant binding for INH, complexes were simulated for 50 ns. It was observed that after simulation, INH remained bound in the pocket with an increased molecular bonding network with the neighbor amino acid residues. In vitro studies clearly suggested that M. smegmatis expressing mutant has a better survival rate in isoniazid treatment as compared to wild type. Conclusion: Overall, this study at the outset suggested that the mutation observed in drug resistant strain provides stability to the Rv1592c protein and increased affinity towards the INH due to flap displacement, leading to the possibility for its modification. In vitro results supported our in silico findings.

No abstract