A Convolution Based Computational Approach Towards DNA N6-methyladenine Site Identification and Motif Extraction in Rice Genome

Abstract

<jats:title>ABSTRACT</jats:title><jats:p>DNA N6-methylation (6mA) in Adenine nucleotide is a post replication modification and is responsible for many biological functions. Experimental methods for genome wide 6mA site detection is an expensive and manual labour intensive process. Automated and accurate computational methods can help to identify 6mA sites in long genomes saving significant time and money. Our study develops a convolutional neural network based tool i6mA-CNN capable of identifying 6mA sites in the rice genome. Our model coordinates among multiple types of features such as PseAAC inspired customized feature vector, multiple one hot representations and dinucleotide physicochemical properties. It achieves area under the receiver operating characteristic curve of 0.98 with an overall accuracy of 0.94 using 5 fold cross validation on benchmark dataset. Finally, we evaluate our model on two other plant genome 6mA site identification datasets besides rice. Results suggest that our proposed tool is able to generalize its ability of 6mA site identification on plant genomes irrespective of plant species. Supplementary data (benchmark dataset, independent test dataset, comparison purpose dataset, trained model, physicochemical property values, attention mechanism details for motif finding) are available at Nucleic Acids Research online. Web tool for this research can be found here.</jats:p>

ACNN-6mA Prediction of N6-Methyladenine Loci in Multiple Species Based on Rice Dataset Pre-training Model†

“…I6mA-Fuse 24 constructs five Random Forest models using five separate feature encoding methods: one-hot, di-nucleotide binary, k-space spectral nucleus, k-mer, and EIIPs, and uses linear regression models to combine the prediction probability scores of five RF models based on a single encoding to predict the m6A sites of F. vesca and R. chinensis. I6mA-CNN 25 uses one-hot, dinucleotide binary, dinucleotide property and a hybrid encoding method to encode the sequence, and then conducts convolution and full join operations to obtain four models and perform fusion to predict methylation sites. Deep6mA 26 encodes the sequence as one pot, extracts, and abstracts the features of the sequence using a 5-layer convolutional neural network, and then uses LSTM to model the extracted abstract features in terms of time series, extracts the timing effect between high-level features, and then makes a full connection prediction.…”