eHSCPr

Hematopoietic stem cells (HSCs) give rise to all blood cells and plays vital roles throughout the whole lifespan through the life cycle through their pluripotency and self-renewal properties. Accurately identifying the stages of early hematopoietic development is extremely important for life sciences and clinical treatment, as it may open up new prospects for extracorporeal blood research. To improve the feature representation, we compared F-score with three state-of-art differential gene selection methods (limma, DESeq2, edgeR) and evaluated their performance on Support Vector Machines (SVM) to determine the optimal gene set. It was observed that the evaluation of the testing dataset shows that F-score outperforms the other three popular methods, achieving the high with area under the curve of receiver operating characteristic (ROC) values of 0.987. The 10-fold cross-validation results based on the optimal gene set indicate that, the SVM has the best robustness and the prediction accuracy is 94.19%, when compared with XGBoost, neural network, and random forest. Additionally, the results obtained from independent dataset demonstrated that ‘ss’ can identify early stages of HSCs development accurately.