Identification of D Modification Sites Using a Random Forest Model Based on Nucleotide Chemical Properties

Abstract

Dihydrouridine (D) is an abundant post-transcriptional modification present in transfer RNA from eukaryotes, bacteria, and archaea. D has contributed to treatments for cancerous diseases. Therefore, the precise detection of D modification sites can enable further understanding of its functional roles. Traditional experimental techniques to identify D are laborious and time-consuming. In addition, there are few computational tools for such analysis. In this study, we utilized eleven sequence-derived feature extraction methods and implemented five popular machine algorithms to identify an optimal model. During data preprocessing, data were partitioned for training and testing. Oversampling was also adopted to reduce the effect of the imbalance between positive and negative samples. The best-performing model was obtained through a combination of random forest and nucleotide chemical property modeling. The optimized model presented high sensitivity and specificity values of 0.9688 and 0.9706 in independent tests, respectively. Our proposed model surpassed published tools in independent tests. Furthermore, a series of validations across several aspects was conducted in order to demonstrate the robustness and reliability of our model.

Long extrachromosomal circular DNA identification by fusing sequence-derived features of physicochemical properties and nucleotide distribution patterns

“…This section summarizes 11 remaining encoders namely, nucleic acid composition (NAC) 62 , enhanced nucleic acid composition (ENAC) 63 , accumulated nucleotide frequency (ANF) 64 , dinucleotide composition (DNC) 65 , trinucleotide composition (TNC) 66 , nucleotide chemical property (NCP) 67 , binary 68 , electron ionic interaction potential (EIIP) 57 , series correlation pseudo dinucleotide composition (SCPseDNC), 69 , pseudo dinucleotide composition (PSEDNC) 70 , 71 , and pseudo k-tupler composition (PSEKNC) 72 .…”

“…Similarly, dinucleotide composition (DNC) 65 and trinucleotide composition (TNC) 66 , use the pairs of nucleotides (k = 2, or k = 3) to compute normalized occurrence frequencies rather than taking into account individual nucleotides. Enhanced nucleic acid composition (ENAC) 63 transforms raw sequences into statistical vectors by counting the number of different k-mers at a fixed sliding window. First, a dictionary of unique k-mers is created and then for each unique each k-mer, within each window its count is computed.…”

iLEC-DNA: Identifying Long Extra-chromosomal Circular DNA by Fusing Sequence-derived Features of Physicochemical Properties and Nucleotide Distribution Patterns

“…Similarly, dinucleotide composition (DNC) 46 and trinucleotide composition (TNC) 47 , use the pairs of nucleotides (k=2, or k=3) to compute normalized occurrence frequencies rather than taking into account individual nucleotides. Enhanced nucleic acid composition (ENAC) 44 transforms raw sequences into statistical vectors by counting the number of different k-mers at a fixed sliding window. First, a dictionary of unique k-mers is created and then for each unique each k-mer, within each window its count is computed.…”

Self-attention enabled deep learning of dihydrouridine (D) modification on mRNAs unveiled a distinct sequence signature from tRNAs

“… 25 A number of computational methods have been developed for predicting epigenetic modifications of RNA. 26 , 27 , 28 , 29 , 30 , 31 , 32 , 33 , 34 , 35 , 36 , 37 Among them, iRNAD is the first approach for D-site prediction from multiple species, which used a support vector machine to distinguish D and non-D sites. 34 Later, iRNAD_XGBoost used XGBoost-selected multiple features to construct a model for D detection.…”

“… 34 Later, iRNAD_XGBoost used XGBoost-selected multiple features to construct a model for D detection. 33 However, to the best of our knowledge, all existing D-site prediction tools 32 , 33 , 34 , 35 were trained on tRNAs, and it is not clear whether they can be applied to predict D sites on mRNAs. Although recent studies have unveiled the widely occurring nature and transcriptome-wide distribution of D (or the D epitranscriptome), 14 there are still no prediction tools constructed for mRNA D sites using mRNA D datasets.…”