Iterative feature representations improve N4-methylcytosine site prediction

Abstract

Motivation Accurate identification of N4-methylcytosine (4mC) modifications in a genome wide can provide insights into their biological functions and mechanisms. Machine learning recently have become effective approaches for computational identification of 4mC sites in genome. Unfortunately, existing methods cannot achieve satisfactory performance, owing to the lack of effective DNA feature representations that are capable to capture the characteristics of 4mC modifications. Results In this work, we developed a new predictor named 4mcPred-IFL, aiming to identify 4mC sites. To represent and capture discriminative features, we proposed an iterative feature representation algorithm that enables to learn informative features from several sequential models in a supervised iterative mode. Our analysis results showed that the feature representations learnt by our algorithm can capture the discriminative distribution characteristics between 4mC sites and non-4mC sites, enlarging the decision margin between the positives and negatives in feature space. Additionally, by evaluating and comparing our predictor with the state-of-the-art predictors on benchmark datasets, we demonstrate that our predictor can identify 4mC sites more accurately. Availability and implementation The user-friendly webserver that implements the proposed 4mcPred-IFL is well established, and is freely accessible at http://server.malab.cn/4mcPred-IFL. Supplementary information Supplementary data are available at Bioinformatics online.

A Deep Neural Network for Identifying DNA N4-Methylcytosine Sites

“…Finally, we get the new C. elegans dataset with 17, 808 samples which contains 111, 73 positive samples and 663, 5 negative samples. The positive samples are the sequences centroided with functional 4mC sites detected by the SMRT sequencing technology, while the negative samples are the sequences with the cytosines in the center but not detected as 4mC (Wei et al, 2019). The new dataset can be downloaded from our github, and the download link is given in section 3.…”

“…For performance evaluation, we used the following five generallyused metrics: Sensitivity (SN), Specificity (SP), Accuracy (ACC), Mathew's Correlation Coefficient (MCC) (Wei et al, 2019) and Area Under the ROC Curve (AUC). The definition of each evaluation metric is as follows: where TP indicates that the actual result is a positive sample, and the predicted result is also a positive sample; TN indicates that the actual result is a negative sample, and the predicted result is also a negative sample; FP indicates that the actual result is a negative sample, and the predicted result is a positive sample (indicating that the negative sample is predicted incorrectly); FN indicates that the actual result is a positive sample, and the prediction result is a negative sample (indicating that the positive sample is predicted incorrectly).…”

“…However, the performance of the experiment did not improve much. Recently, Wei et al (2019) proposed the fourth 4mC site predictor, called 4mcPred-IFL, which uses an iterative feature representation algorithm to learn probabilistic features from different sequential models and enhance feature representation in a supervised iterative manner. However, the complexity of 4mcPred-IFL is very high.…”

“…We use the eight features mentioned inChen et al (2017),He et al (2018),Wei et al (2018), andWei et al (2019) as preliminary features. These features are obtained by encoding the different sequence information by the feature representation algorithm of the sequence.…”

“…These features are BKF (Binary and k-mer frequency), DBPF (Dinucleotide binary profile and frequency), KNN (K-Nearest Neighbor), PCP (Physical-Chemical Properties), MMI (Multivariate Mutual Information), PseDNC (Pseudo dinucleotide composition), PseEIIP (Electronion interaction pseudopotentials of trinucleotide) and RFHCP (Ring-function-hydrogen-chemical properties). The related feature extraction methods can be found inWei et al (2019).…”