Filters








20,201 Hits in 5.5 sec

MODEL: Motif-Based Deep Feature Learning for Link Prediction

Lei Wang, Jing Ren, Bo Xu, Jianxin Li, Wei Luo, Feng Xia
2020 IEEE Transactions on Computational Social Systems  
Recently, approaches for link prediction have evolved from traditional similarity-based algorithms into embedding-based algorithms.  ...  To evaluate its effectiveness for link prediction, experiments were conducted on three types of networks: social networks, biological networks, and academic networks.  ...  In this paper, we propose a MOtif-based DEep feature learning algorithm for Link prediction named MODEL. The proposed algorithm uses motifs to automatically learn vectorized features.  ... 
doi:10.1109/tcss.2019.2962819 fatcat:e2t5pzsu5zayhlawv5fe7j3lie

RelEx: A Model-Agnostic Relational Model Explainer [article]

Yue Zhang, David Defazio, Arti Ramesh
2020 arXiv   pre-print
This is essential, as complex deep learning models with millions of parameters produce state of the art results, but it can be nearly impossible to explain their predictions.  ...  In recent years, considerable progress has been made on improving the interpretability of machine learning models.  ...  For example, the pixels that are instrumental in a prediction is not often apparent for state-of-the-art deep neural network models for many vision tasks [3] .  ... 
arXiv:2006.00305v1 fatcat:roqfurwfrng4tdq2akzovltawi

Discovering epistatic feature interactions from neural network models of regulatory DNA sequences [article]

Peyton G Greenside, Tyler Shimko, Polly Fordyce, Anshul Kundaje
2018 bioRxiv   pre-print
Several feature attribution methods have been developed for estimating the predictive importance of individual features (nucleotides or motifs) in any input DNA sequence to its associated output prediction  ...  Our approach makes significant strides in improving the interpretability of deep learning models for genomics. Availability: Code is available at: https://github.com/kundajelab/dfim  ...  Acknowledgements We would like to thank Johnny Israeli and Nathan Boley for their help training a TAL1, GATA1, GATA2 transcription factor binding model.  ... 
doi:10.1101/302711 fatcat:khjgdo5fnrcdzmchvuzx45twdq

Attention please: modeling global and local context in glycan structure-function relationships [article]

Bowen Dai, Daniel E Mattox, Chris Bailey-Kellogg
2021 bioRxiv   pre-print
Our method, glyBERT, encodes glycans with a branched biochemical language and employs an attention-based deep language model to learn biologically relevant glycan representations focused on the most important  ...  Applying glyBERT to a variety of prediction tasks confirms the value of capturing rich context-dependent patterns in this attention-based model: the same monosaccharides and glycan motifs are represented  ...  School of Engineering, and their lab members, for their continued feedback on this project and many others at all stages of development.  ... 
doi:10.1101/2021.10.15.464532 fatcat:h4kjctak4zfu3gjstgemceogqu

LectinOracle - A Generalizable Deep Learning Model for Lectin-Glycan Binding Prediction [article]

Jon Lundstrøm, Emma Korhonen, Frederique LISACEK, Daniel Bojar
2021 bioRxiv   pre-print
Here, we present LectinOracle, a model combining transformer-based representations for proteins and graph convolutional neural networks for glycans to predict their interaction.  ...  Yet the sheer breadth and depth of specificity for diverse oligosaccharide motifs has made studying lectins a largely piecemeal approach, with few options to generalize.  ...  In general, a deep learning-based glycan analysis module learns similarities between glycan motifs and can more easily generalize to new motifs than a discretized, motif counting-based approach that is  ... 
doi:10.1101/2021.08.30.458147 fatcat:k57iz7b3t5gt3gydubqafsqgum

Discovering epistatic feature interactions from neural network models of regulatory DNA sequences

Peyton Greenside, Tyler Shimko, Polly Fordyce, Anshul Kundaje
2018 Bioinformatics  
Our approach makes significant strides in improving the interpretability of deep learning models for genomics.  ...  Several feature attribution methods have been developed for estimating the predictive importance of individual features (nucleotides or motifs) in any input DNA sequence to its associated output prediction  ...  Acknowledgements We would like to thank Johnny Israeli and Nathan Boley for their help training a TAL1, GATA1, GATA2 transcription factor binding model.  ... 
doi:10.1093/bioinformatics/bty575 pmid:30423062 pmcid:PMC6129272 fatcat:mhuwtfitqjhnbl4zbesup2ixly

PRPI-SC: an ensemble deep learning model for predicting plant lncRNA-protein interactions

Haoran Zhou, Jael Sanyanda Wekesa, Yushi Luan, Jun Meng
2021 BMC Bioinformatics  
Results In this study, we propose an ensemble deep learning model to predict plant lncRNA-protein interactions using stacked denoising autoencoder and convolutional neural network based on sequence and  ...  PRPI-SC predicts interactions between lncRNAs and proteins based on the k-mer features of RNAs and proteins.  ...  the deep learning model.  ... 
doi:10.1186/s12859-021-04328-9 fatcat:xg6busdvqfay7pxvmfqludqkvu

A deep learning framework for modeling structural features of RNA-binding protein targets

Sai Zhang, Jingtian Zhou, Hailin Hu, Haipeng Gong, Ligong Chen, Chao Cheng, Jianyang Zeng
2015 Nucleic Acids Research  
In this paper, we develop a general and flexible deep learning framework for modeling structural binding preferences and predicting binding sites of RBPs, which takes (predicted) RNA tertiary structural  ...  In addition, we have conducted the first study to show that integrating the additional RNA tertiary structural features can improve the model performance in predicting RBP binding sites, especially for  ...  Hu for the insightful discussions and assistance on the docking experiments, and Mr H.C. Gong for his help on the figure preparation.  ... 
doi:10.1093/nar/gkv1025 pmid:26467480 pmcid:PMC4770198 fatcat:jxntktgumbcdhiodbwim6v24cm

Explainability in transformer models for functional genomics

Jim Clauwaert, Gerben Menschaert, Willem Waegeman
2021 Briefings in Bioinformatics  
The effectiveness of deep learning methods can be largely attributed to the automated extraction of relevant features from raw data.  ...  This work builds upon a transformer-based neural network framework designed for prokaryotic genome annotation purposes.  ...  Introduction Deep learning techniques are increasingly obtaining state-ofthe-art performances for a multitude of prediction tasks in genomics [8] .  ... 
doi:10.1093/bib/bbab060 pmid:33834200 pmcid:PMC8425421 fatcat:ikaqsgin6jepjatltpw23u447a

Chromatin interaction aware gene regulatory modeling with graph attention networks

Alireza Karbalayghareh, Merve Sahin, Christina S Leslie
2022 Genome Research  
levels than state-of-the-art deep learning methods for this task.  ...  Sequence-based GraphReg also accurately predicts direct transcription factor (TF) targets as validated by CRISPRi TF knockout experiments via in silico ablation of TF binding motifs.  ...  Author contributions: A.K. developed the machine learning models, performed all computational experiments, and co-wrote the manuscript. M.S. performed analysis on HiChIP data sets.  ... 
doi:10.1101/gr.275870.121 pmid:35396274 pmcid:PMC9104700 fatcat:nytcvidiprgylmzqgtcslaomvi

Harnessing Current Knowledge of DNA N6-Methyladenosine From Model Plants for Non-model Crops

Sadaruddin Chachar, Jingrong Liu, Pingxian Zhang, Adeel Riaz, Changfei Guan, Shuyuan Liu
2021 Frontiers in Genetics  
We also consider advanced artificial-intelligence biotechnologies that improve extraction and prediction of 6mA concepts.  ...  In this Review, we discuss the potential challenges that may hinder exploitation of 6mA, and give future goals of 6mA from model plants to non-model crops.  ...  In addition, as deep learning approaches have become a powerful strategy for modeling and prediction "Big Data" in genomics and epigenomics (including 6mA), from our perspective, it will lead to detailed  ... 
doi:10.3389/fgene.2021.668317 pmid:33995495 pmcid:PMC8118384 fatcat:xz2wvd6syjaktg2yl62ydkza44

Thermodynamic modeling reveals widespread multivalent binding by RNA-binding proteins [article]

Salma Sohrabi-Jahromi, Johannes Soeding
2021 bioRxiv   pre-print
Results: We present Bipartite Motif Finder (BMF), which is based on a thermodynamic model of RBPs with two cooperatively binding RNA-binding domains.  ...  Our results demonstrate the importance of multivalent binding for RNA-binding proteins and highlight the value of bipartite motif models in representing the multivalency of protein-RNA interactions.  ...  Acknowledgements We thank Christian Roth for his help on AVXoptimization, web server implementation and for discussions on tool development and benchmark design.  ... 
doi:10.1101/2021.01.30.428941 fatcat:e5v4hlqdyfcb5g6z3vfhy5xpsq

Network representation learning: models, methods and applications

Anuraj Mohan, K. V. Pramod
2019 SN Applied Sciences  
Definition 5 A signed network is a network G = (V , E) , v ∈ V , e ∈ E and for each edge, e ij = +1 or e ij = −1 , denoting a positive link or a negative link between v i and v j .  ...  We discuss the common models of network representation learning and reviews the major works which come under each model with respect to each type of network.  ...  Acknowledgements The authors would like to thank the management and staff of Department of Computer Applications, CUSAT, India and NSS College of Engineering, Palakkad, India for providing enough materials  ... 
doi:10.1007/s42452-019-1044-9 fatcat:zvlbj4qozzfw3dxoyevb6wgska

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

Qinhu Zhang, Zhen Shen, De-Shuang Huang
2019 Scientific Reports  
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  ...  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  ...  Soon after, a number of deep-learning based methods are proposed for modeling in-vivo protein-DNA binding [13] [14] [15] [16] [17] .  ... 
doi:10.1038/s41598-019-44966-x pmid:31186519 pmcid:PMC6559991 fatcat:aru5ewgl5jf7rlwzbpqf3ttbsm

Towards More Realistic Simulated Datasets for Benchmarking Deep Learning Models in Regulatory Genomics [article]

Eva Prakash, Avanti Shrikumar, Anshul Kundaje
2021 bioRxiv   pre-print
These models are appealing in part because they can learn predictive DNA sequence features without prior assumptions.  ...  Our analysis suggests several promising directions for future research on these model interpretation methods. Code and links to data are available at https://github.com/kundajelab/interpret-benchmark.  ...  a strong baseline alongside deep learning models (Fig. 2 ).  ... 
doi:10.1101/2021.12.26.474224 fatcat:sbi7m7cjgjf2tpxlerkm7ceasu
« Previous Showing results 1 — 15 out of 20,201 results