A deep learning framework combined with word embedding to identify DNA replication origins

Abstract

The DNA replication influences the inheritance of genetic information in the DNA life cycle. As the distribution of replication origins (ORIs) is the major determinant to precisely regulate the replication process, the correct identification of ORIs is significant in giving an insightful understanding of DNA replication mechanisms and the regulatory mechanisms of genetic expressions. For eukaryotes in particular, multiple ORIs exist in each of their gene sequences to complete the replication in a reasonable period of time. To simplify the identification process of eukaryote's ORIs, most of existing methods are developed by traditional machine learning algorithms, and target to the gene sequences with a fixed length. Consequently, the identification results are not satisfying, i.e. there is still great room for improvement. To break through the limitations in previous studies, this paper develops sequence segmentation methods, and employs the word embedding technique, 'Word2vec', to convert gene sequences into word vectors, thereby grasping the inner correlations of gene sequences with different lengths. Then, a deep learning framework to perform the ORI identification task is constructed by a convolutional neural network with an embedding layer. On the basis of the analysis of similarity reduction dimensionality diagram, Word2vec can effectively transform the inner relationship among words into numerical feature. For four species in this study, the best models are obtained with the overall accuracy of 0.975, 0.765, 0.885, 0.967, the Matthew's correlation coefficient of 0.940, 0.530, 0.771, 0.934, and the AUC of 0.975, 0.800, 0.888, 0.981, which indicate that the proposed predictor has a stable ability and provide a high confidence coefficient to classify both of ORIs and non-ORIs. Compared with state-of-the-art methods, the proposed predictor can achieve ORI identification with significant improvement. It is therefore reasonable to anticipate that the proposed method will make a useful high throughput tool for genome analysis.

Figure 1

The overall workflow of the proposed method.

Figure 2

The 2-dimensional feature space generated by the t-SNE algorithm in dataset $S_1$ using Continuous-TSSS. (a) The first training result. (b) The second training result. (c) The third training result.

Figure 3

The 2-dimensional feature space generated by the t-SNE algorithm in the dataset $S_1$ Using Skip-TSSS. (a) The first training result. (b) The second training result. (c) The third training result.

Figure 4

ROC curves obtained by using different training modes in each dataset based on Continuous-TSSS. (a) ROC curves of $S_1$. (b) ROC curves of $S_2$. (c) ROC curves of $S_3$. (d) ROC curves of $S_4$.

Figure 5

ROC curves obtained by using different training modes in each dataset based on Skip-TSSS. (a) ROC curves of $S_1$. (b) ROC curves of $S_2$. (c) ROC curves of $S_3$. (d) ROC curves of $S_4$.

Figure 6

Nucleotide compositions in the flanking sequences of eukaryotic ORIs and bacterial ORIs. (a) Average mononucleotide frequencies in the flanking sequences of eukaryotic ORIs and bacterial ORIs. (b) Average dinucleotide frequencies in the flanking sequences of eukaryotic ORIs and bacterial ORIs. (c) Average trinucleotide frequencies in the flanking sequences of eukaryotic ORIs and bacterial ORIs.

Figure 7

The process of Skip-three slices sequence segmentation (Skip-TSSS).

Figure 8

The architecture of the proposed CNN.

Figure 9

The un-trainable embedding layer construction.

Figure 10

The trainable embedding layer construction.

Figure 11

The two-channel embedding layer construction.

Figure 12

The sizes of the convolutional kernel employed in the CNN.