Table. 1 Parse enhancer syntax with machine learning algorithms.
From: Enhancer reprogramming: critical roles in cancer and promising therapeutic strategies
Classification | Algorithms | Description | References |
|---|---|---|---|
Support Vector Machine (SVM) | ChromaGenSVM | Predicting enhancers based on the optimal combination of histone epigenetic markers. | [350] |
Gkm-SVM | Use SVM to identify enhancer regions by calculating nucleotide level importance scores. | [351] | |
EnhancerFinder | Multi-kernel learning algorithm based on support vector machine for predicting developmental enhancers. | [352] | |
Random Forest (RF) | RFECS | Based on the random forest algorithm, predict enhancers by analyzing chromatin state data. | [353] |
DRAF | Used to predict TFBS. | [354] | |
Convolutional Neural Network (CNN) | DeepEnhancer | Predicting enhancers through deep convolutional neural networks and processing variable length sequences in the FANTOM5 dataset. | [355] |
DeepBind | Used to predict TFBS. | [356] | |
DESSO | Used to predict regulatory motifs from human ChIP-seq data. | [357] | |
Basset | Predicting chromatin accessibility by classifying features in the sequence. | [358] | |
DeepSTARR | Used to predict enhancers with development activities. | [359] | |
BPNet | Predicting transcription factor binding profiles at single nucleotide resolution using convolutional neural networks. | [360] | |
Enformer | Predicting enhancer-promoter interactions from DNA sequences. | [60] | |
iEnhancer-ECNN | Using the ensemble of CNN to identify enhancers and their intensities. | [361] | |
GhmCN | Predicting gene expression status 5hmC signaling and prioritizing potential enhancers | [362] | |
DeepCAPE | Predicting enhancers by integrating DNA sequences and DNase-seq data. | [363] | |
CoNSEPT | Used to predict enhancer function under different conditions, such as cell type and experimental conditions. | [364] | |
DEEPSEN | Integrating 36 features for predicting super-enhancers. | [365] | |
ES-ARCNN | Expand the previously identified enhancer dataset through data augmentation techniques (i.e. reverse complement and shift). | [366] | |
Hybrid architecture | DanQ | Combining CNN and bidirectional recurrent neural networks for predicting the features of regulatory regions. | [367] |
BiRen | Combining CNN and bidirectional recurrent neural networks for predicting enhancers. | [368] | |
DeepATT | A mixed-class attention neural network for identifying functional effects of DNA sequences. | [369] | |
Enhancer-IF | Integrating RF, extremely random tree, multilayer perceptron, SVM, and extreme gradient enhancement to enhance model robustness. | [370] | |
SEMet | Predicting enhancers associated with metastasis and prognosis of pancreatic cancer by SE landscape. | [371] | |
CAPReSE | Using CNN and XGBoost to identify specific structural variation (SV)-mediated aberrant chromatin contacts in cancer genomes, with particular application in colorectal cancer. | [107] |
