Publications

2024

1. PNAS

   How to improve polygenic prediction from increasingly prevalent whole-genome sequencing data?

   Wanwen Zeng, Hanmin Guo, Qiao Liu, and Wing H. Wong^†

   Proceedings of the National Academy of Sciences, 2024

   under review

   Abs HTML PDF

   Polygenic risk scores (PRS) are crucial in genetics for predicting individual susceptibility to complex diseases by aggregating the effects of numerous genetic variants. Whole-genome sequencing (WGS) has revolutionized our ability to detect rare and even de novo variants, creating an exciting opportunity for developing new PRS methods that can effectively leverage rare variants and capture the complex relationships among different variants. Furthermore, regulatory mechanisms play a crucial role in gene expression and disease manifestation, offering avenues to further enhance the performance and interpretation of PRS predictions. Through simulation studies, we highlighted aspects where current PRS methods face challenges when applied to WGS data, aiming to shed light on potential opportunities for further improvement. To address these challenges, we developed Epi-PRS, an approach that leverages the power of genomic large language models (LLM) to impute epigenomic signals across diverse cellular contexts, for use as intermediate variables between genotype and phenotype. A pretrained LLM is employed to transform genotypes into epigenomic signals using personal diploid sequences as inputs, and the genetic risk is then estimated based on the imputed personal epigenomic signals. Epi-PRS enhances the assessment of personal variant impacts, enabling a comprehensive and holistic consideration of genotypic and regulatory information within large genomic regions. Our simulation results demonstrated that incorporating the nuanced effects of non-linear models, rare variants, and regulatory information can provide more precise PRS prediction and better understanding of genetic risk. Applying Epi-PRS to real data from the UK Biobank, our results further showed that Epi-PRS significantly outperforms existing PRS methods in two major diseases: breast cancer and diabetes. This study suggests that PRS methods can benefit from incorporating non-linear models, rare variants, and regulatory information, highlighting the potential for significant advancements in disease risk modeling and enhancing the understanding of precision medicine.

2. Nat. Aging

   Associating genotype to imaging and clinical phenotypes of Alzheimer’s disease by leveraging genomic large language model

   Qiao Liu^*, Wanwen Zeng^*, Hongtu Zhu, Lexin Li^†, and Wing H. Wong^†

   Nature Aging, 2024

   under review

   Abs HTML PDF

   Genome-wide association studies (GWAS) have identified numerous genetic variants associated with Alzheimer’s disease (AD) phenotypes. However, how these variants contribute to the etiology of AD remains largely elusive. Recent advances in genomic large language models (LLMs) have revolutionized regulatory genomic prediction tasks, offering new opportunities to interpret the genetic variation observed in personal genome. In this study, we propose epiBrainLLM, a novel computational framework that leverages genomic LLM to enhance our understanding of the causal pathways from genotypes to brain measures to AD-related clinical phenotypes. Our framework will first convert the personal DNA sequence into a diverse set of genomic and epigenomic features using a pretrained genomic LLM and then use these features to further predict phenotypes. Across various experimental settings, our results demonstrate that incorporating pretrained genomic LLMs significantly improves association analysis compared to using genotype information alone. We conclude that our proposed framework provides a novel perspective for understanding the regulatory mechanisms underlying the AD disease etiology, potentially offering insights into complex disease mechanisms beyond AD.

3. Nat. Commun.

   CREATE: cell-type-specific cis-regulatory elements identification via discrete embedding

   Xuejian Cui, Qijin Yin, Zijing Gao, Zhen Li, Xiaoyang Chen, Shengquan Chen, Qiao Liu, Wanwen Zeng^†, and Rui Jiang^†

   Nature Communications, 2024

   in revision

   Abs HTML PDF

   Identifying cis-regulatory elements (CREs) within non-coding genomic regions-such as enhancers, silencers, promoters, and insulators-is pivotal for elucidating the intricate gene regulatory mechanisms underlying complex biological traits. The current prevalent sequence-based methods often focus on singular CRE types, limiting insights into cell-type-specific biological implications. Here, we introduce CREATE, a multimodal deep learning model based on the Vector Quantized Variational AutoEncoder framework, designed to extract discrete CRE embeddings and classify multiple CRE classes using genomic sequences, chromatin accessibility, and chromatin interaction data. CREATE excels in accurate CRE identification and exhibits strong effectiveness and robustness. We showcase CREATE’s capability in generating comprehensive CRE-specific feature spectrum, offering quantitative and interpretable insights into CRE specificity. By enabling large-scale prediction of CREs in specific cell types, CREATE facilitates the recognition of disease- or phenotype-related biological variabilities of CREs, thereby expanding our understanding of gene regulation landscapes.

4. RECOMB

   bpBERT: base-resolution BERT models reveal DNA sequence regulatory syntax and variants

   Wanwen Zeng, Shuang Chen, Yuti Liu, and Wing H. Wong^†

   2024

   Under review

   Abs HTML PDF

   bpBERT introduces a BERT-based model to predict DNA regulatory sequences and identify variants with base-level precision.

5. BIBM

   Inferring gene regulatory networks based on genetically perturbed scATAC-seq data

   Wei Shao, Shuang Zhang, Qiao Liu, and Wanwen Zeng^†

   IEEE International Conference on Bioinformatics and Biomedicine, 2024

   Abs HTML PDF

   Gene regulatory networks (GRNs) are critical blueprints for understanding gene regulation and the intricate interactions that drive biological processes. Recent advances have highlighted the potential of single-cell ATAC-seq (scATAC-seq) data in GRN inference, offering unprecedented insights into how chromatin accessibility plays an important part in gene regulation. However, existing methods often fall short in providing a quantitative and holistic depiction of regulatory relationships, particularly in capturing the strength, direction, and type of gene regulation simultaneously. In this paper, we present a novel approach that addresses these limitations by leveraging genetically perturbed scATAC-seq data to infer more comprehensive and accurate GRNs. Our method advances the field by integrating pre- and post-perturbation chromatin accessibility data, enabling the construction of GRNs that more accurately reflect the dynamic regulatory landscape. Through rigorous evaluation on seven real datasets, we demonstrate the method’s superior performance in reconstructing GRNs with enhanced precision and interpretability. This work significantly contributes to the field by providing a robust framework for GRN inference, with broad implications for understanding gene regulation in complex biological systems.

6. Genome Biol.

   EpiGePT: a Pretrained Transformer model for epigenomics

   Zijing Gao^*, Qiao Liu^*†, Wanwen Zeng, Rui Jiang^†, and Wing Hung Wong^†

   Genome Biology, 2024

   minor revision

   Abs HTML PDF

   The transformer-based models, such as GPT-31 and DALL-E2, have achieved unprecedented breakthroughs in the field of natural language processing and computer vision. The inherent similarities between natural language and biological sequences have prompted a new wave of inferring the grammatical rules underneath the biological sequences. In genomic study, it is worth noting that DNA sequences alone cannot explain all the gene activities due to epigenetic mechanism. To investigate this problem, we propose EpiGePT, a new transformer-based language pretrained model in epigenomics, for predicting genome-wide epigenomic signals by considering the mechanistic modeling of transcriptional regulation. Specifically, EpiGePT takes the context-specific activities of transcription factors (TFs) into consideration, which could offer deeper biological insights comparing to models trained on DNA sequence only. In a series of experiments, EpiGePT demonstrates state-of-the-art performance in a diverse epigenomic signals prediction tasks as well as new prediction tasks by fine-tuning. Furthermore, EpiGePT is capable of learning the cell-type-specific long-range interactions through the self-attention mechanism and interpreting the genetic variants that associated with human diseases. We expect that the advances of EpiGePT can shed light on understanding the complex regulatory mechanisms in gene regulation. We provide free online prediction service of EpiGePT through https://health.tsinghua.edu.cn/epigept/.

7. Cyber. and Intell.

   Modeling the causal mechanism between genotypes and phenotypes using large-scale biobank data and context-specific regulatory networks

   Wenran Li^*, Wanwen Zeng^*, and Wing H. Wong^†

   Cybernetics and Intelligence, 2024

   Abs HTML PDF

   The relationship between genetic variation and human phenotypes is crucial for developing effective treatments and personalized medicine. However, our understanding of the regulatory mechanisms by which variants influence human traits and diseases is far from complete. Context-specific regulatory network is a typical tool that provides detailed understanding of gene regulation in specific biological contexts, allowing us to identify key regulators and pathways that are important for a particular phenotype. In this review, we summarize the large international biobanks and reference omics data that provide diverse datasets for the genotypephenotype analysis and the construction of context-specific regulatory networks, and discuss the importance of context-specific regulatory networks in explaining the underlying causal mechanism between genotypes and phenotypes. We emphasize the significance of QTL studies in explaining the correlation between genotypes and omics features, and present various computational approaches for the construction of context-specific regulatory networks. With continued advancements in biobanking, genomics, and computational biology, the context-specific regulatory networks may serve as an increasingly powerful tool for modeling the causal mechanisms that underlie the relationship between genotypes and phenotypes.

2023

1. QB

   Deepdrug: A general graph‐based deep learning framework for drug‐drug interactions and drug‐target interactions prediction

   Qijin Yin, Rui Fan, Xusheng Cao, Qiao Liu^†, Rui Jiang^†, and Wanwen Zeng^†

   Quantitative Biology, 2023

   Abs HTML PDF

   In this paper, we develop DeepDrug, a deep learning framework for overcoming the above limitation by using residual graph convolutional networks (Res-GCNs) and convolutional networks (CNNs) to learn the comprehensive structure- and sequence-based representations of drugs and proteins.

2. Bioinform. Adv.

   Applications of transformer-based language models in bioinformatics: a survey

   Shuang Zhang, Shuang Chen, Yuti Liu, Qiao Liu, and Wanwen Zeng^†

   Bioinformatics Advances, 2023

   Abs HTML PDF

   The transformer-based language models, including vanilla transformer, BERT and GPT-3, have achieved revolutionary breakthroughs in the field of natural language processing (NLP). Since there are inherent similarities between various biological sequences and natural languages, the remarkable interpretability and adaptability of these models have prompted a new wave of their application in bioinformatics research. To provide a timely and comprehensive review, we introduce key developments of transformer-based language models by describing the detailed structure of transformers and summarize their contribution to a wide range of bioinformatics research from basic sequence analysis to drug discovery. While transformer-based applications in bioinformatics are diverse and multifaceted, we identify and discuss the common challenges, including heterogeneity of training data, computational expense and model interpretability, and opportunities in the context of bioinformatics research. We hope that the broader community of NLP researchers, bioinformaticians and biologists will be brought together to foster future research and development in transformer-based language models, and inspire novel bioinformatics applications that are unattainable by traditional methods.

3. Brief. Bioinform.

   Deep generative modeling and clustering of single-cell Hi-C data

   Qiao Liu^*, Wanwen Zeng^*, Wei Zhang, Sicheng Wang, Hongyang Chen, Rui Jiang^†, Mu Zhou^†, and Shaoting Zhang^†

   Briefings in Bioinformatics, 2023

   Abs HTML PDF

   Deciphering 3D genome conformation is important for understanding gene regulation and cellular function at a spatial level. The recent advances of single cell Hi-C technologies have enabled the profiling of the 3D architecture of DNA within individual cell, which allows us to study the cell-to-cell variability of 3D chromatin organization. Computational approaches are in urgent need to comprehensively analyze the sparse and heterogeneous single cell Hi-C data. Here, we proposed scDEC-Hi-C, a new framework for single cell Hi-C analysis with deep generative neural networks. scDEC-Hi-C outperforms existing methods in terms of single cell Hi-C data clustering and imputation. Moreover, the generative power of scDEC-Hi-C could help unveil the differences of chromatin architecture across cell types. We expect that scDEC-Hi-C could shed light on deepening our understanding of the complex mechanism underlying the formation of chromatin contacts.

2022

1. NAR

   HiChIPdb: A database of HiChIP regulatory interactions

   Wanwen Zeng^*, Qiao Liu^*, Qijin Yin^*, Rui Jiang^†, and Wing H. Wong^†

   Nucleic Acids Research, 2022

   Abs HTML PDF

   Elucidating the role of 3D architecture of DNA in gene regulation is crucial for understanding cell differentiation, tissue homeostasis and disease development. Among various chromatin conformation capture methods, HiChIP has received increasing attention for its significant improvement over other methods in profiling of regulatory (e.g. H3K27ac) and structural (e.g. cohesin) interactions. To facilitate the studies of 3D regulatory interactions, we developed a HiChIP interactions database, HiChIPdb (http://health.tsinghua.edu.cn/hichipdb/). The current version of HiChIPdb contains ∼262M annotated HiChIP interactions from 200 high-throughput HiChIP samples across 108 cell types. The functionalities of HiChIPdb include: (i) standardized categorization of HiChIP interactions in a hierarchical structure based on organ, tissue and cell line and (ii) comprehensive annotations of HiChIP interactions with regulatory genes and GWAS Catalog SNPs. To the best of our knowledge, HiChIPdb is the first comprehensive database that utilizes a unified pipeline to map the functional interactions across diverse cell types and tissues in different resolutions. We believe this database has the potential to advance cutting-edge research in regulatory mechanisms in development and disease by removing the barrier in data aggregation, preprocessing, and analysis.

2. Bioinfo.

   scGraph: A graph neural network-based approach to automatically identify cell types

   Qijin Yin, Qiao Liu, Zhuoran Fu, Wanwen Zeng, Boheng Zhang, Xuegong Zhang, Rui Jiang^†, and Hairong Lv^†

   Bioinformatics, 2022

   Abs HTML PDF

   Single-cell technologies play a crucial role in revolutionizing biological research over the past decade, which strengthens our understanding in cell differentiation, development and regulation from a single-cell level perspective. Single-cell RNA sequencing (scRNA-seq) is one of the most common single cell technologies, which enables probing transcriptional states in thousands of cells in one experiment. Identification of cell types from scRNA-seq measurements is a fundamental and crucial question to answer. Most previous studies directly take gene expression as input while ignoring the comprehensive gene–gene interactions. We propose scGraph, an automatic cell identification algorithm leveraging gene interaction relationships to enhance the performance of the cell-type identification. scGraph is based on a graph neural network to aggregate the information of interacting genes. In a series of experiments, we demonstrate that scGraph is accurate and outperforms eight comparison methods in the task of cell-type identification. Moreover, scGraph automatically learns the gene interaction relationships from biological data and the pathway enrichment analysis shows consistent findings with previous analysis, providing insights on the analysis of regulatory mechanism.

2021

1. Nat. Mach. Intell.

   Reusability report: Compressing regulatory networks to vectors for interpreting gene expression and genetic variants

   Wanwen Zeng^*, Jingxue Xin^*, Rui Jiang^†, and Yong Wang^†

   Nature Machine Intelligence, 2021

   Abs HTML PDF

   Overall, we have demonstrated the robustness and re-usability of the GEEK embedding framework and shown its convenience for follow-up studies by representing the network as a vector. After systematic evaluation of GEEK and other applications, we expect that the incorporation of sequence-based features and 3D chromatin interaction-based features into a unified framework may provide a holistic perspective to understand gene transcriptional control and potentially provide insights to identify genome-wide association study risk variants. In addition, multidimensional summaries of omics data can be further integrated into sophisticated statistical models. For example, the STAAR model introduces ‘annotation principal components’ to effectively summarize multiple qualitative and quantitative variant functional annotations to boost the power of variant set tests for continuous and binary traits in whole-genome sequencing rare variants association studies.

2. NAR

   SilencerDB: A comprehensive database of silencers

   Wanwen Zeng^*, Shengquan Chen^*, Xuejian Cui^*, Xiaoyang Chen, Zijing Gao, and Rui Jiang^†

   Nucleic Acids Research, 2021

   Abs HTML PDF

   Gene regulatory elements, including promoters, enhancers, silencers, etc., control transcriptional programs in a spatiotemporal manner. Though these elements are known to be able to induce either positive or negative transcriptional control, the community has been mostly studying enhancers which amplify transcription initiation, with less emphasis given to silencers which repress gene expression. To facilitate the study of silencers and the investigation of their potential roles in transcriptional control, we developed SilencerDB (http://health.tsinghua.edu.cn/silencerdb/), a comprehensive database of silencers by manually curating silencers from 2300 published articles. The current version, SilencerDB 1.0, contains (1) 33 060 validated silencers from experimental methods, and (ii) 5 045 547 predicted silencers from state-of-the-art machine learning methods. The functionality of SilencerDB includes (a) standardized categorization of silencers in a tree-structured class hierarchy based on species, organ, tissue and cell line and (b) comprehensive annotations of silencers with the nearest gene and potential regulatory genes. SilencerDB, to the best of our knowledge, is the first comprehensive database at this scale dedicated to silencers, with reliable annotations and user-friendly interactive database features. We believe this database has the potential to enable advanced understanding of silencers in regulatory mechanisms and to empower researchers to devise diverse applications of silencers in disease development.

2020

1. Bioinformatics

   Integrating distal and proximal information to predict gene expression via a densely connected convolutional neural network

   Wanwen Zeng, Yong Wang, and Rui Jiang^†

   Bioinformatics, 2020

   Abs HTML PDF

   We propose DeepExpression, a densely connected convolutional neural network, to predict gene expression using both promoter sequences and enhancer–promoter interactions. We demonstrate that our model consistently outperforms baseline methods, not only in the classification of binary gene expression status but also in regression of continuous gene expression levels, in both cross-validation experiments and cross-cell line predictions. We show that the sequential promoter information is more informative than the experimental enhancer information; meanwhile, the enhancer–promoter interactions within ±100 kbp around the TSS of a gene are most beneficial. We finally visualize motifs in both promoter and enhancer regions and show the match of identified sequence signatures with known motifs. We expect to see a wide spectrum of applications using HiChIP data in deciphering the mechanism of gene regulation.

2. Quant. Biol.

   IRIS: A method for predicting in vivo RNA secondary structures using PARIS data

   Jianyu Zhou, Pan Li, Wanwen Zeng, Wenxiu Ma, Zhipeng Lu, Rui Jiang, Qiangfeng Cliff Zhang, and Tao Jiang^†

   Quantitative Biology, 2020

   Abs HTML PDF

   We introduce IRIS, a method for predicting RNA secondary structure ensembles based on PARIS data. IRIS generates a large set of candidate RNA secondary structure models under the guidance of redistributed PARIS reads and then uses a Bayesian model to identify the optimal ensemble, according to both thermodynamic principles and PARIS data.

2019

1. Nat. Commun.

   DC3: A method for deconvolution and coupled clustering from bulk and single-cell genomics data

   Wanwen Zeng^†, Xi Chen^†, Zhana Duren^†, Yong Wang, Rui Jiang^†, and Wing Hung Wong^†

   Nature Communications, 2019

   Abs HTML PDF

   Characterizing and interpreting heterogeneous mixtures at the cellular level is a critical problem in genomics. Single-cell assays offer an opportunity to resolve cellular level heterogeneity, e.g., scRNA-seq enables single-cell expression profiling, and scATAC-seq identifies active regulatory elements. Furthermore, while scHi-C can measure the chromatin contacts (i.e., loops) between active regulatory elements to target genes in single cells, bulk HiChIP can measure such contacts in a higher resolution. In this work, we introduce DC3 (De-Convolution and Coupled-Clustering) as a method for the joint analysis of various bulk and single-cell data such as HiChIP, RNA-seq and ATAC-seq from the same heterogeneous cell population. DC3 can simultaneously identify distinct subpopulations, assign single cells to the subpopulations (i.e., clustering) and de-convolve the bulk data into subpopulation-specific data. The subpopulation-specific profiles of gene expression, chromatin accessibility and enhancer-promoter contact obtained by DC3 provide a comprehensive characterization of the gene regulatory system in each subpopulation.

2. Database

   EnDisease: A manually curated database for enhancer-disease associations

   Wanwen Zeng, Xu Min, and Rui Jiang^†

   Database, 2019

   Abs HTML PDF

   Genome-wide association studies have successfully identified thousands of genomic loci potentially associated with hundreds of complex traits in the past decade. Nevertheless, the fact that more than 90% of such disease-associated variants lie in non-coding DNA with unknown functional implications has been appealing for advanced analysis of plenty of genetic variants. Toward this goal, recent studies focusing on individual non-coding variants have revealed that complex diseases are often the consequences of erroneous interactions between enhancers and their target genes. However, such enhancer-disease associations are dispersed in a variety of independent studies, and thus far it is still difficult to carry out comprehensive downstream analysis with these experimentally supported enhancer-disease associations. To fill in this gap, we collected experimentally supported associations between complex diseases and enhancers and then developed a manually curated database called EnDisease (http://bioinfo.au.tsinghua.edu.cn/endisease/). Concretely, EnDisease documents 535 associations between 133 diseases and 454 enhancers, extracted from 199 articles. Moreover, after annotating these enhancers using 649 human and 115 mouse DNase-seq experiments, we find that cancer-related enhancers tend to be open across a large number of cell types. This database provides a user-friendly interface for browsing and searching, and it also allows users to download data freely. EnDisease has the potential to become a helpful and important resource for researchers who aim to understand the molecular mechanisms of enhancers involved in complex diseases.

2018

1. PNAS

   Integrative analysis of single-cell genomics data by coupled nonnegative matrix factorizations

   Zhana Duren^*, Xi Chen^*, Mahdi Zamanighomi^*, Wanwen Zeng, Ansuman T. Satpathy, Howard Y. Chang, Yong Wang^†, and Wing Hung Wong^†

   Proceedings of the National Academy of Sciences, 2018

   Abs HTML PDF

   When different types of functional genomics data are generated on single cells from different samples of cells from the same heterogeneous population, the clustering of cells in the different samples should be coupled. We formulate this “coupled clustering” problem as an optimization problem and propose the method of coupled nonnegative matrix factorizations (coupled NMF) for its solution. The method is illustrated by the integrative analysis of single-cell RNA-sequencing (RNA-seq) and single-cell ATAC-sequencing (ATAC-seq) data.

2. Methods

   Leveraging multiple gene networks to prioritize GWAS candidate genes via network representation learning

   Mengmeng Wu, Wanwen Zeng, Wenqiang Liu, Hairong Lv, Ting Chen, and Rui Jiang^†

   Methods, 2018

   Abs HTML PDF

   Genome-wide association studies (GWAS) have successfully discovered a number of disease-associated genetic variants in the past decade, providing an unprecedented opportunity for deciphering genetic basis of human inherited diseases. However, it is still a challenging task to extract biological knowledge from the GWAS data, due to such issues as missing heritability and weak interpretability. Indeed, the fact that the majority of discovered loci fall into noncoding regions without clear links to genes has been preventing the characterization of their functions and appealing for a sophisticated approach to bridge genetic and genomic studies. Towards this problem, network-based prioritization of candidate genes, which performs integrated analysis of gene networks with GWAS data, has emerged as a promising direction and attracted much attention. However, most existing methods overlook the sparse and noisy properties of gene networks and thus may lead to suboptimal performance. Motivated by this understanding, we proposed a novel method called REGENT for integrating multiple gene networks with GWAS data to prioritize candidate genes for complex diseases. We leveraged a technique called the network representation learning to embed a gene network into a compact and robust feature space, and then designed a hierarchical statistical model to integrate features of multiple gene networks with GWAS data for the effective inference of genes associated with a disease of interest. We applied our method to six complex diseases and demonstrated the superior performance of REGENT over existing approaches in recovering known disease-associated genes. We further conducted a pathway analysis and showed that the ability of REGENT to discover disease-associated pathways. We expect to see applications of our method to a broad spectrum of diseases for post-GWAS analysis. REGENT is freely available at https://github.com/wmmthu/REGENT.

2017

1. BMC Bioinfo.

   Predicting enhancers with deep convolutional neural networks

   Xu Min^*, Wanwen Zeng^*, Shengquan Chen, Ning Chen, Ting Chen, and Rui Jiang^†

   BMC Bioinformatics, 2017

   Abs HTML PDF

   To facilitate the identification of enhancers, we propose a computational framework, named DeepEnhancer, to distinguish enhancers from background genomic sequences. Our method purely relies on DNA sequences to predict enhancers in an end-to-end manner by using a deep convolutional neural network (CNN). We train our deep learning model on permissive enhancers and then adopt a transfer learning strategy to fine-tune the model on enhancers specific to a cell line. Results demonstrate the effectiveness and efficiency of our method in the classification of enhancers against random sequences, exhibiting advantages of deep learning over traditional sequence-based classifiers. We then construct a variety of neural networks with different architectures and show the usefulness of such techniques as max-pooling and batch normalization in our method. To gain the interpretability of our approach, we further visualize convolutional kernels as sequence logos and successfully identify similar motifs in the JASPAR database.

2. Bioinformatics

   Chromatin accessibility prediction via convolutional long short-term memory networks with k-mer embedding

   Xu Min, Wanwen Zeng, Ning Chen, Ting Chen, and Rui Jiang^†

   Bioinformatics, 2017

   Abs HTML PDF

   We fill this gap by addressing the problem of chromatin accessibility prediction with a convolutional Long Short-Term Memory (LSTM) network with k-mer embedding. We first split DNA sequences into k-mers and pre-train k-mer embedding vectors based on the co-occurrence matrix of k-mers by using an unsupervised representation learning approach. We then construct a supervised deep learning architecture comprised of an embedding layer, three convolutional layers and a Bidirectional LSTM (BLSTM) layer for feature learning and classification. We demonstrate that our method gains high-quality fixed-length features from variable-length sequences and consistently outperforms baseline methods. We show that k-mer embedding can effectively enhance model performance by exploring different embedding strategies. We also prove the efficacy of both the convolution and the BLSTM layers by comparing two variations of the network architecture. We confirm the robustness of our model to hyper-parameters by performing sensitivity analysis. We hope our method can eventually reinforce our understanding of employing deep learning in genomic studies and shed light on research regarding mechanisms of chromatin accessibility.

3. BMC Syst. Biol.

   Mimvec: A deep learning approach for analyzing the human phenome

   Mingxin Gan, Wenran Li, Wanwen Zeng, Xiaojian Wang, and Rui Jiang^†

   BMC Systems Biology, 2017

   Abs HTML PDF

   Our method Mimvec is capable of not only capturing semantic relationships between words in biomedical records but also alleviating the dimensional disaster accompanying the traditional TF-IDF framework. With the approaching of precision medicine, there will be abundant electronic records of medicine and health awaiting for deep analysis, and we expect to see a wide spectrum of applications borrowing the idea of our method in the near future.