Modeling in-vivo protein-DNA binding by combining multiple-instance learning with a hybrid deep neural network

Abstract

Modeling in-vivo protein-DNA binding is not only fundamental for further understanding of the regulatory mechanisms, but also a challenging task in computational biology. Deep-learning based methods have succeed in modeling in-vivo protein-DNA binding, but they often (1) follow the fully supervised learning framework and overlook the weakly supervised information of genomic sequences that a bound DNA sequence may has multiple TFBS(s), and, (2) use one-hot encoding to encode DNA sequences and ignore the dependencies among nucleotides. In this paper, we propose a weakly supervised framework, which combines multiple-instance learning with a hybrid deep neural network and uses k -mer encoding to transform DNA sequences, for modeling in-vivo protein-DNA binding. Firstly, this framework segments sequences into multiple overlapping instances using a sliding window, and then encodes all instances into image-like inputs of high-order dependencies using k -mer encoding. Secondly, it separately computes a score for all instances in the same bag using a hybrid deep neural network that integrates convolutional and recurrent neural networks. Finally, it integrates the predicted values of all instances as the final prediction of this bag using the Noisy-and method. The experimental results on in-vivo datasets demonstrate the superior performance of the proposed framework. In addition, we also explore the performance of the proposed framework when using k -mer encoding, and demonstrate the performance of the Noisy-and method by comparing it with other fusion methods, and find that adding recurrent layers can improve the performance of the proposed framework.

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

“…It uses the noisy method to integrate the prediction values of all instances into the final prediction of the package. The improved method is termed as WSCNNLSTM [89].…”

“…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.…”

“…Scholars have, therefore, proposed hybrid methods for predicting TFBSs in complex sequence data. Examples include combination models such as CNN + LSTM [89], sequence2vec + CNN [83], word2vec + GRU [87], and attention + CNN [70]. Using the combination models, researchers can identify TFBSs in complex sequences, thereby allowing for a better understanding of the transcription process.…”

A Novel Neural Model With Lateral Interaction for Learning Tasks

“…Neural models have been rapidly developed since the MP model (McCulloch & Pitts, 1943) was proposed to mode the structure of human brain. Simulating learning activities for engineering applications is always an important goal in the development of neural models (see Snyder et al, 2016;Han et al, 2016;Zhang, Zhen, & Huang, 2019). Generally, learning activities can be considered as supervised learning and unsupervised learning.…”

Assessing deep learning algorithms incis-regulatory motif finding based on genomic sequencing data

“…To address this problem, we benchmarked the existing 20 DL tools [9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] ( Supplementary Table S1) using 690 ENCODE DNA ChIP-Seq datasets (covering 91 TFs in 161 cell lines) and 55 RNA CLIP-Seq datasets (Fig. 1A-B and ( Supplementary Table S2) 28,23 .…”