RESUMO
The effectiveness and side effects of Type 2 diabetes (T2D) medication are related to individual genetic background. SNPs CYP3A4 and CYP2C19 were introduced to machine-learning models to improve the performance of T2D medication prediction. Two multilabel classification models, ML-KNN and WRank-SVM, trained with clinical data and CYP3A4/CYP2C19 SNPs were evaluated. Prediction performance was evaluated with Hamming loss, one-error, coverage, ranking loss and average precision. The average precision of ML-KNN and WRank-SVM using clinical data was 92.74% and 92.9%, respectively. Combined with CYP2C19*2*3, the average precision dropped to 88.84% and 89.93%, respectively. While combined with CYP3A4*1G, the average precision was enhanced to 97.96% and 97.82%, respectively. Results suggest that CYP3A4*1G can improve the performance of ML-KNN and WRank-SVM models in predicting T2D medication performance.
About 10% of adults around the world are living with Type 2 Diabetes (T2D). Due to the huge number of patients and the complexity of individual makeup, it is a challenge for doctors to prescribe appropriate hypoglycemic drugs. To aid prescribing, machine-learning models were developed to predict medication schemes based on patients' demographic information and laboratory test results. These models treat prediction as a multilabel classification problem, with each class of medication as a label. This work was designed to determine whether the introduction of genetic information would improve prediction performance. The machine-learning models were trained using datasets with and without genetic information and their performance was compared. The performance of the machine-learning models was improved by incorporating the SNP CYP3A4*1G into the datasets. Thus, this work demonstrates a novel strategy to improve the prediction of T2D hypoglycemic medication performance and provides new ideas for how to support the T2D health system with machine-learning techniques.
