Deep Motif Dashboard: Visualizing and Understanding Genomic Sequences Using Deep Neural Networks

Abstract

Deep neural network (DNN) models have recently obtained state-of-the-art prediction accuracy for the transcription factor binding (TFBS) site classification task. However, it remains unclear how these approaches identify meaningful DNA sequence signals and give insights as to why TFs bind to certain locations. In this paper, we propose a toolkit called the Deep Motif Dashboard (DeMo Dashboard) which provides a suite of visualization strategies to extract motifs, or sequence patterns from deep neural network models for TFBS classification. We demonstrate how to visualize and understand three important DNN models: convolutional, recurrent, and convolutional-recurrent networks. Our first visualization method is finding a test sequence’s saliency map which uses first-order derivatives to describe the importance of each nucleotide in making the final prediction. Second, considering recurrent models make predictions in a temporal manner (from one end of a TFBS sequence to the other), we introduce temporal output scores, indicating the prediction score of a model over time for a sequential input. Lastly, a class-specific visualization strategy finds the optimal input sequence for a given TFBS positive class via stochastic gradient optimization. Our experimental results indicate that a convolutional-recurrent architecture performs the best among the three architectures. The visualization techniques indicate that CNN-RNN makes predictions by modeling both motifs as well as dependencies among them.

GaKCo: A Fast Gapped k-mer String Kernel Using Counting

“…This is because GaKCo's gapped k-mer formulation is the same as gkm-SVM but with an improved (faster) implementation. Besides, in the supplementary, we also compare GaKCo's empirical performance with a state-of-the-art CNN model [17]. For 16/19 tasks, GaKCo outperforms the CNN model with an average of ∼ 20% improvements.…”

“…We also compare GaKCo to the CNN implementation from [17] for all the datasets (results in supplementary). Classification: After calculation, we input the N × N kernel matrix into an SVM classifier as an empirical feature map using a linear kernel in LIBLINEAR [10].…”

“…Later, Alipanahi et al [1] applied a convolutional neural network model for predicting sequence specificity of DNA and RNA-binding proteins as well as generating motifs, or consensus patterns, from the features that were learned by their model. Lanchantin et al [17] proposed a deep convolutional/highway MLP framework for the same task and demonstrated improved performance. In the field of natural language processing, multiple works like [30] have used deep learning models for document [33] or sentiment [29] Transcription factors (TFs) are regulatory proteins that bind to functional sites of DNA to control the regulation of genes.…”

“…Recently, Deep Neural Networks (NNs) have provided state-of-the-art performances for various sequence classification tasks like analyzing DNA [1,17], proteins [25,34], and natural language [33,29] sequences. Despite their superior performance in accuracy and speed (e.g.…”

“…Recently, Deep Neural Networks (NNs) have provided state-of-the-art performances for various sequence classification tasks. We compare GaKCo's empirical performance with a state-of-the-art deep convolutional neural network (CNN) model [17]. On datasets with few training samples, GaKCo achieves an average accuracy improvement of 20% over the CNN model (details in the supplementary).…”