Multi-task Deep Neural Networks in Automated Protein Function Prediction

Background: In recent years, deep learning algorithms have outperformed the state-of-the art methods in several areas such as computer vision, speech recognition thanks to the efficient methods for training and for preventing overfitting, advancement in computer hardware and the availability of vast amount data. The high performance of multi-task deep neural networks in drug discovery has attracted the attention to deep learning algorithms in the bioinformatics area. Protein function prediction is a crucial research area where more accurate prediction methods are still needed. Here, we proposed a hierarchical multi-task deep neural network architecture based on Gene Ontology (GO) terms as a solution to the protein function prediction problem and investigated various aspects of the proposed architecture by performing several experiments. ∗A short version of this manuscript was accepted for oral presentation at ISMB/ECCB 2017 Function-COSI meeting †[email protected] ‡[email protected] §[email protected] ¶[email protected] ‖[email protected] 1 ar X iv :1 70 5. 04 80 2v 2 [ qbi o. Q M ] 2 8 M ay 2 01 7 Results: First, we showed that there is a positive correlation between the performance of the system and the size of training datasets. Second, we investigated whether the level of GO terms on the GO hierarchy is related to their performance. We showed that there is no relation between the depth of GO terms on the GO hierarchy (i.e. general/specific) and their performance. In addition, we included all annotations to the training of a set of GO terms to investigate whether including noisy data to the training datasets change the performance of the system. The results showed that including less reliable annotations in training of deep neural networks increased the performance of the low performed GO terms, significantly. Finally, we evaluated the performance of the system using hierarchical evaluation method. Mathews correlation coefficients was calculated as 0.75, 0.49 and 0.63 for molecular function, biological process and cellular component categories, respectively. Conclusions: We showed that deep learning algorithms have a great potential in protein function prediction area. We plan to further improve the DEEPred by including other types of annotations from various biological data sources. Finally, we plan to construct DEEPred as an open access online tool.