ABSTRACTEssential genes are necessary to the survival or reproduction of a living organism. The prediction and analysis of gene essentiality can advance our understanding to basic life and human diseases, and further boost the development of new drugs. Wet lab methods for identifying essential genes are often costly, time consuming, and laborious. As a complement, computational methods have been proposed to predict essential genes by integrating multiple biological data sources. Most of these methods are evaluated on model organisms. However, prediction methods for human essential genes are still limited and the relationship between human gene essentiality and different biological information still needs to be explored. In addition, exploring suitable deep learning techniques to overcome the limitations of traditional machine learning methods and improve the prediction accuracy is also important and interesting. We propose a deep learning based method, DeepSF, to predict human essential genes. DeepSF integrates sequence features derived from DNA and protein sequence data with features extracted or learned from different types of functional data, such as gene ontology, protein complex, protein domain, and protein-protein interaction network. More than 200 features from these biological data are extracted/learned which are integrated together to train a cost-sensitive deep neural network by utilizing multiple deep leaning techniques. The experimental results of 10-fold cross validation show that DeepSF can accurately predict human gene essentiality with an average AUC of 95.17%, the area under precision-recall curve (auPRC) of 92.21%, the accuracy of 91.59%, and the F1 measure about 78.71%. In addition, the comparison experimental results show that DeepSF significantly outperforms several popular traditional machine learning models (SVM, Random Forest, and Adaboost), and performs slightly better than a recent deep learning model (DeepHE). We have demonstrated that the proposed method, DeepSF, is effective for predicting human essential genes. Deep learning techniques are promising at both feature learning and classification levels for the task of essential gene prediction.
Cross-Predicting Essential Genes between Two Model Eukaryotic Species Using Machine Learning
