{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/deep-motif-dashboard-visualizing-and","title":"Deep Motif Dashboard: Visualizing and Understanding Genomic Sequences Using Deep Neural Networks","arxiv_id":"1608.03644","date":"2016-08-12","proceeding":null,"authors":["Jack Lanchantin","Ritambhara Singh","Beilun Wang","Yanjun Qi"],"abstract":"Deep neural network (DNN) models have recently obtained state-of-the-art\nprediction accuracy for the transcription factor binding (TFBS) site\nclassification task. However, it remains unclear how these approaches identify\nmeaningful DNA sequence signals and give insights as to why TFs bind to certain\nlocations. In this paper, we propose a toolkit called the Deep Motif Dashboard\n(DeMo Dashboard) which provides a suite of visualization strategies to extract\nmotifs, or sequence patterns from deep neural network models for TFBS\nclassification. We demonstrate how to visualize and understand three important\nDNN models: convolutional, recurrent, and convolutional-recurrent networks. Our\nfirst visualization method is finding a test sequence's saliency map which uses\nfirst-order derivatives to describe the importance of each nucleotide in making\nthe final prediction. Second, considering recurrent models make predictions in\na temporal manner (from one end of a TFBS sequence to the other), we introduce\ntemporal output scores, indicating the prediction score of a model over time\nfor a sequential input. Lastly, a class-specific visualization strategy finds\nthe optimal input sequence for a given TFBS positive class via stochastic\ngradient optimization. Our experimental results indicate that a\nconvolutional-recurrent architecture performs the best among the three\narchitectures. The visualization techniques indicate that CNN-RNN makes\npredictions by modeling both motifs as well as dependencies among them.","url_abs":"http://arxiv.org/abs/1608.03644v4","url_pdf":"http://arxiv.org/pdf/1608.03644v4.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"deep-motif-dashboard-visualizing-and","repo_url":"https://github.com/QData/DeepMotif","is_official":0,"mentioned_in_paper":0,"mentioned_in_github":1,"framework":"torch","reach":null}],"tasks":[{"task_slug":"classification","task_name":"General Classification"}],"methods":[],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}