Sequence-Based Prediction of Putative Transcription Factor Binding Sites in DNA Sequences of Any Length

Abstract

A transcription factor (TF) is a protein that regulates gene expression by binding to specific DNA sequences. Despite the recent advances in experimental techniques for identifying transcription factor binding sites (TFBS) in DNA sequences, a large number of TFBS are to be unveiled in many species. Several computational methods developed for predicting TFBS in DNA are tissue- or species-specific methods, so cannot be used without prior knowledge of tissue or species. Some computational methods are applicable to finding TFBS in short DNA sequences only. In this paper we propose a new learning method for predicting TFBS in DNA of any length using the composition, transition and distribution of nucleotides and amino acids in DNA and TF sequences. In independent testing of the method on datasets that were not used in training the method, its accuracy and MCC were as high as 81.84% and 0.634, respectively. The proposed method can be a useful aid for selecting potential TFBS in a large amount of DNA sequences before conducting biochemical experiments to empirically determine TFBS. The program and data sets are available at http://bclab.inha.ac.kr/TFbinding.

Recent advances in exploring transcriptional regulatory landscape of crops

“…Machine learning-based approaches establish predictive criteria by learning from documented TF-TFBS data using diverse computational strategies. For instance, Lee and coworkers introduced a SVM model incorporating various features of TFs and TFBSs, achieving approximately 82% prediction accuracy ( Lee et al., 2017c ). A recent study achieved a remarkable 99% accuracy in model prediction by integrating the chemical properties of TF proteins, along with the structural conformation and bonding capabilities of both TFs and DNA ( Khamis et al., 2018 ).…”

A Review About Transcription Factor Binding Sites Prediction Based on Deep Learning

“…In recent years, prediction of TFBSs based on sequence data has been achieved in some studies. Among these, the methods to determine the presence of motifs are mainly based on traditional machine learning methods such as EM and support-vector machines [99], [100], and deep learning methods such as CNN and LSTM [24], [89]. Due to the abstract and sequential nature of DNA sequences, prediction method of TFBSs solely based on sequence is not suitable for complex sequence data.…”

Plant-DTI: Extending the landscape of TF protein and DNA interaction in plants by a machine learning-based approach

“… Qian et al (2006) modeled the TF recognition sequences from binary-based patterns of short TFBS motifs and the TFs with corresponding gene ontology using the k-nearest neighbor classifier, with approximately 77% model accuracy. Lee et al (2017) later introduced an SVM model with more features of TFs and TFBS properties, improving the accuracy of prediction to 82%. The model considered the global composition of residues in TFs and TFBSs sequences using the composition, transition, and distribution of residues in the sequences.…”