Loading…
DeepDiff: DEEP-learning for predicting DIFFerential gene expression from histone modifications
Abstract Motivation Computational methods that predict differential gene expression from histone modification signals are highly desirable for understanding how histone modifications control the functional heterogeneity of cells through influencing differential gene regulation. Recent studies either...
Saved in:
Published in: | Bioinformatics 2018-09, Vol.34 (17), p.i891-i900 |
---|---|
Main Authors: | , , |
Format: | Article |
Language: | English |
Citations: | Items that this one cites Items that cite this one |
Online Access: | Request full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
cited_by | cdi_FETCH-LOGICAL-c463t-cdc2e850a9a9be124ef1c8a26ae4b907440a59b21bb397671c85459c878f48c43 |
---|---|
cites | cdi_FETCH-LOGICAL-c463t-cdc2e850a9a9be124ef1c8a26ae4b907440a59b21bb397671c85459c878f48c43 |
container_end_page | i900 |
container_issue | 17 |
container_start_page | i891 |
container_title | Bioinformatics |
container_volume | 34 |
creator | Sekhon, Arshdeep Singh, Ritambhara Qi, Yanjun |
description | Abstract
Motivation
Computational methods that predict differential gene expression from histone modification signals are highly desirable for understanding how histone modifications control the functional heterogeneity of cells through influencing differential gene regulation. Recent studies either failed to capture combinatorial effects on differential prediction or primarily only focused on cell type-specific analysis. In this paper we develop a novel attention-based deep learning architecture, DeepDiff, that provides a unified and end-to-end solution to model and to interpret how dependencies among histone modifications control the differential patterns of gene regulation. DeepDiff uses a hierarchy of multiple Long Short-Term Memory (LSTM) modules to encode the spatial structure of input signals and to model how various histone modifications cooperate automatically. We introduce and train two levels of attention jointly with the target prediction, enabling DeepDiff to attend differentially to relevant modifications and to locate important genome positions for each modification. Additionally, DeepDiff introduces a novel deep-learning based multi-task formulation to use the cell-type-specific gene expression predictions as auxiliary tasks, encouraging richer feature embeddings in our primary task of differential expression prediction.
Results
Using data from Roadmap Epigenomics Project (REMC) for ten different pairs of cell types, we show that DeepDiff significantly outperforms the state-of-the-art baselines for differential gene expression prediction. The learned attention weights are validated by observations from previous studies about how epigenetic mechanisms connect to differential gene expression.
Availability and implementation
Codes and results are available at deepchrome.org.
Supplementary information
Supplementary data are available at Bioinformatics online. |
doi_str_mv | 10.1093/bioinformatics/bty612 |
format | article |
fullrecord | <record><control><sourceid>proquest_TOX</sourceid><recordid>TN_cdi_proquest_miscellaneous_2133437910</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><oup_id>10.1093/bioinformatics/bty612</oup_id><sourcerecordid>2133437910</sourcerecordid><originalsourceid>FETCH-LOGICAL-c463t-cdc2e850a9a9be124ef1c8a26ae4b907440a59b21bb397671c85459c878f48c43</originalsourceid><addsrcrecordid>eNqNUMlOwzAQtRCIlsIngHLkErBjZzE31LRQqRIc4ErkuONilNjBTiT697hKQeLGaZa3zOghdEnwDcGc3tbaaqOsa0Wvpb-t-11GkiM0JSzDcYJTfhx6muUxKzCdoDPvPzBOCWPsFE0oZgnFeTZFbyVAV2ql7qJysXiOGxDOaLONgnXUOdho2e_HcrVcggPTa9FEWzAQwVeAvdfWRMrZNnrXvrdh39qNVlqGt6zx5-hEicbDxaHO0Oty8TJ_jNdPD6v5_TqWLKN9LDcygSLFggteA0kYKCILkWQCWM1xzhgWKa8TUteU51kewJSlXBZ5oVghGZ2h69G3c_ZzAN9XrfYSmkYYsIOvEkIpozknOFDTkSqd9d6BqjqnW-F2FcHVPtrqb7TVGG3QXR1ODHULm1_VT5aBgEeCHbp_en4DIwKMWg</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2133437910</pqid></control><display><type>article</type><title>DeepDiff: DEEP-learning for predicting DIFFerential gene expression from histone modifications</title><source>OUP_牛津大学出版社OA刊</source><creator>Sekhon, Arshdeep ; Singh, Ritambhara ; Qi, Yanjun</creator><creatorcontrib>Sekhon, Arshdeep ; Singh, Ritambhara ; Qi, Yanjun</creatorcontrib><description>Abstract
Motivation
Computational methods that predict differential gene expression from histone modification signals are highly desirable for understanding how histone modifications control the functional heterogeneity of cells through influencing differential gene regulation. Recent studies either failed to capture combinatorial effects on differential prediction or primarily only focused on cell type-specific analysis. In this paper we develop a novel attention-based deep learning architecture, DeepDiff, that provides a unified and end-to-end solution to model and to interpret how dependencies among histone modifications control the differential patterns of gene regulation. DeepDiff uses a hierarchy of multiple Long Short-Term Memory (LSTM) modules to encode the spatial structure of input signals and to model how various histone modifications cooperate automatically. We introduce and train two levels of attention jointly with the target prediction, enabling DeepDiff to attend differentially to relevant modifications and to locate important genome positions for each modification. Additionally, DeepDiff introduces a novel deep-learning based multi-task formulation to use the cell-type-specific gene expression predictions as auxiliary tasks, encouraging richer feature embeddings in our primary task of differential expression prediction.
Results
Using data from Roadmap Epigenomics Project (REMC) for ten different pairs of cell types, we show that DeepDiff significantly outperforms the state-of-the-art baselines for differential gene expression prediction. The learned attention weights are validated by observations from previous studies about how epigenetic mechanisms connect to differential gene expression.
Availability and implementation
Codes and results are available at deepchrome.org.
Supplementary information
Supplementary data are available at Bioinformatics online.</description><identifier>ISSN: 1367-4803</identifier><identifier>EISSN: 1460-2059</identifier><identifier>EISSN: 1367-4811</identifier><identifier>DOI: 10.1093/bioinformatics/bty612</identifier><identifier>PMID: 30423076</identifier><language>eng</language><publisher>England: Oxford University Press</publisher><ispartof>Bioinformatics, 2018-09, Vol.34 (17), p.i891-i900</ispartof><rights>The Author(s) 2018. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com 2018</rights><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c463t-cdc2e850a9a9be124ef1c8a26ae4b907440a59b21bb397671c85459c878f48c43</citedby><cites>FETCH-LOGICAL-c463t-cdc2e850a9a9be124ef1c8a26ae4b907440a59b21bb397671c85459c878f48c43</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,780,784,1604,27924,27925</link.rule.ids><linktorsrc>$$Uhttps://dx.doi.org/10.1093/bioinformatics/bty612$$EView_record_in_Oxford_University_Press$$FView_record_in_$$GOxford_University_Press</linktorsrc><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/30423076$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Sekhon, Arshdeep</creatorcontrib><creatorcontrib>Singh, Ritambhara</creatorcontrib><creatorcontrib>Qi, Yanjun</creatorcontrib><title>DeepDiff: DEEP-learning for predicting DIFFerential gene expression from histone modifications</title><title>Bioinformatics</title><addtitle>Bioinformatics</addtitle><description>Abstract
Motivation
Computational methods that predict differential gene expression from histone modification signals are highly desirable for understanding how histone modifications control the functional heterogeneity of cells through influencing differential gene regulation. Recent studies either failed to capture combinatorial effects on differential prediction or primarily only focused on cell type-specific analysis. In this paper we develop a novel attention-based deep learning architecture, DeepDiff, that provides a unified and end-to-end solution to model and to interpret how dependencies among histone modifications control the differential patterns of gene regulation. DeepDiff uses a hierarchy of multiple Long Short-Term Memory (LSTM) modules to encode the spatial structure of input signals and to model how various histone modifications cooperate automatically. We introduce and train two levels of attention jointly with the target prediction, enabling DeepDiff to attend differentially to relevant modifications and to locate important genome positions for each modification. Additionally, DeepDiff introduces a novel deep-learning based multi-task formulation to use the cell-type-specific gene expression predictions as auxiliary tasks, encouraging richer feature embeddings in our primary task of differential expression prediction.
Results
Using data from Roadmap Epigenomics Project (REMC) for ten different pairs of cell types, we show that DeepDiff significantly outperforms the state-of-the-art baselines for differential gene expression prediction. The learned attention weights are validated by observations from previous studies about how epigenetic mechanisms connect to differential gene expression.
Availability and implementation
Codes and results are available at deepchrome.org.
Supplementary information
Supplementary data are available at Bioinformatics online.</description><issn>1367-4803</issn><issn>1460-2059</issn><issn>1367-4811</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2018</creationdate><recordtype>article</recordtype><recordid>eNqNUMlOwzAQtRCIlsIngHLkErBjZzE31LRQqRIc4ErkuONilNjBTiT697hKQeLGaZa3zOghdEnwDcGc3tbaaqOsa0Wvpb-t-11GkiM0JSzDcYJTfhx6muUxKzCdoDPvPzBOCWPsFE0oZgnFeTZFbyVAV2ql7qJysXiOGxDOaLONgnXUOdho2e_HcrVcggPTa9FEWzAQwVeAvdfWRMrZNnrXvrdh39qNVlqGt6zx5-hEicbDxaHO0Oty8TJ_jNdPD6v5_TqWLKN9LDcygSLFggteA0kYKCILkWQCWM1xzhgWKa8TUteU51kewJSlXBZ5oVghGZ2h69G3c_ZzAN9XrfYSmkYYsIOvEkIpozknOFDTkSqd9d6BqjqnW-F2FcHVPtrqb7TVGG3QXR1ODHULm1_VT5aBgEeCHbp_en4DIwKMWg</recordid><startdate>20180901</startdate><enddate>20180901</enddate><creator>Sekhon, Arshdeep</creator><creator>Singh, Ritambhara</creator><creator>Qi, Yanjun</creator><general>Oxford University Press</general><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>7X8</scope></search><sort><creationdate>20180901</creationdate><title>DeepDiff: DEEP-learning for predicting DIFFerential gene expression from histone modifications</title><author>Sekhon, Arshdeep ; Singh, Ritambhara ; Qi, Yanjun</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c463t-cdc2e850a9a9be124ef1c8a26ae4b907440a59b21bb397671c85459c878f48c43</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2018</creationdate><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Sekhon, Arshdeep</creatorcontrib><creatorcontrib>Singh, Ritambhara</creatorcontrib><creatorcontrib>Qi, Yanjun</creatorcontrib><collection>PubMed</collection><collection>CrossRef</collection><collection>MEDLINE - Academic</collection><jtitle>Bioinformatics</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Sekhon, Arshdeep</au><au>Singh, Ritambhara</au><au>Qi, Yanjun</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>DeepDiff: DEEP-learning for predicting DIFFerential gene expression from histone modifications</atitle><jtitle>Bioinformatics</jtitle><addtitle>Bioinformatics</addtitle><date>2018-09-01</date><risdate>2018</risdate><volume>34</volume><issue>17</issue><spage>i891</spage><epage>i900</epage><pages>i891-i900</pages><issn>1367-4803</issn><eissn>1460-2059</eissn><eissn>1367-4811</eissn><abstract>Abstract
Motivation
Computational methods that predict differential gene expression from histone modification signals are highly desirable for understanding how histone modifications control the functional heterogeneity of cells through influencing differential gene regulation. Recent studies either failed to capture combinatorial effects on differential prediction or primarily only focused on cell type-specific analysis. In this paper we develop a novel attention-based deep learning architecture, DeepDiff, that provides a unified and end-to-end solution to model and to interpret how dependencies among histone modifications control the differential patterns of gene regulation. DeepDiff uses a hierarchy of multiple Long Short-Term Memory (LSTM) modules to encode the spatial structure of input signals and to model how various histone modifications cooperate automatically. We introduce and train two levels of attention jointly with the target prediction, enabling DeepDiff to attend differentially to relevant modifications and to locate important genome positions for each modification. Additionally, DeepDiff introduces a novel deep-learning based multi-task formulation to use the cell-type-specific gene expression predictions as auxiliary tasks, encouraging richer feature embeddings in our primary task of differential expression prediction.
Results
Using data from Roadmap Epigenomics Project (REMC) for ten different pairs of cell types, we show that DeepDiff significantly outperforms the state-of-the-art baselines for differential gene expression prediction. The learned attention weights are validated by observations from previous studies about how epigenetic mechanisms connect to differential gene expression.
Availability and implementation
Codes and results are available at deepchrome.org.
Supplementary information
Supplementary data are available at Bioinformatics online.</abstract><cop>England</cop><pub>Oxford University Press</pub><pmid>30423076</pmid><doi>10.1093/bioinformatics/bty612</doi><oa>free_for_read</oa></addata></record> |
fulltext | fulltext_linktorsrc |
identifier | ISSN: 1367-4803 |
ispartof | Bioinformatics, 2018-09, Vol.34 (17), p.i891-i900 |
issn | 1367-4803 1460-2059 1367-4811 |
language | eng |
recordid | cdi_proquest_miscellaneous_2133437910 |
source | OUP_牛津大学出版社OA刊 |
title | DeepDiff: DEEP-learning for predicting DIFFerential gene expression from histone modifications |
url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T11%3A14%3A51IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_TOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=DeepDiff:%20DEEP-learning%20for%20predicting%20DIFFerential%20gene%20expression%20from%20histone%20modifications&rft.jtitle=Bioinformatics&rft.au=Sekhon,%20Arshdeep&rft.date=2018-09-01&rft.volume=34&rft.issue=17&rft.spage=i891&rft.epage=i900&rft.pages=i891-i900&rft.issn=1367-4803&rft.eissn=1460-2059&rft_id=info:doi/10.1093/bioinformatics/bty612&rft_dat=%3Cproquest_TOX%3E2133437910%3C/proquest_TOX%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c463t-cdc2e850a9a9be124ef1c8a26ae4b907440a59b21bb397671c85459c878f48c43%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=2133437910&rft_id=info:pmid/30423076&rft_oup_id=10.1093/bioinformatics/bty612&rfr_iscdi=true |