MCI-frcnn: A deep learning method for topological micro-domain boundary detection
Chromatin structural domains, or topologically associated domains (TADs), are a general organizing principle in chromatin biology. RNA polymerase II (RNAPII) mediates multiple chromatin interactive loops, tethering together as RNAPII-associated chromatin interaction domains (RAIDs) to offer a framew...
Main Authors: | , , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Frontiers Media S.A.
2022-11-01
|
Series: | Frontiers in Cell and Developmental Biology |
Subjects: | |
Online Access: | https://www.frontiersin.org/articles/10.3389/fcell.2022.1050769/full |
_version_ | 1797986674071830528 |
---|---|
author | Simon Zhongyuan Tian Pengfei Yin Kai Jing Yang Yang Yewen Xu Guangyu Huang Duo Ning Melissa J. Fullwood Melissa J. Fullwood Melissa J. Fullwood Meizhen Zheng |
author_facet | Simon Zhongyuan Tian Pengfei Yin Kai Jing Yang Yang Yewen Xu Guangyu Huang Duo Ning Melissa J. Fullwood Melissa J. Fullwood Melissa J. Fullwood Meizhen Zheng |
author_sort | Simon Zhongyuan Tian |
collection | DOAJ |
description | Chromatin structural domains, or topologically associated domains (TADs), are a general organizing principle in chromatin biology. RNA polymerase II (RNAPII) mediates multiple chromatin interactive loops, tethering together as RNAPII-associated chromatin interaction domains (RAIDs) to offer a framework for gene regulation. RAID and TAD alterations have been found to be associated with diseases. They can be further dissected as micro-domains (micro-TADs and micro-RAIDs) by clustering single-molecule chromatin-interactive complexes from next-generation three-dimensional (3D) genome techniques, such as ChIA-Drop. Currently, there are few tools available for micro-domain boundary identification. In this work, we developed the MCI-frcnn deep learning method to train a Faster Region-based Convolutional Neural Network (Faster R-CNN) for micro-domain boundary detection. At the training phase in MCI-frcnn, 50 images of RAIDs from Drosophila RNAPII ChIA-Drop data, containing 261 micro-RAIDs with ground truth boundaries, were trained for 7 days. Using this well-trained MCI-frcnn, we detected micro-RAID boundaries for the input new images, with a fast speed (5.26 fps), high recognition accuracy (AUROC = 0.85, mAP = 0.69), and high boundary region quantification (genomic IoU = 76%). We further applied MCI-frcnn to detect human micro-TADs boundaries using human GM12878 SPRITE data and obtained a high region quantification score (mean gIoU = 85%). In all, the MCI-frcnn deep learning method which we developed in this work is a general tool for micro-domain boundary detection. |
first_indexed | 2024-04-11T07:36:53Z |
format | Article |
id | doaj.art-dc9b36bedb3f46beb7c5ce9856ed3ac6 |
institution | Directory Open Access Journal |
issn | 2296-634X |
language | English |
last_indexed | 2024-04-11T07:36:53Z |
publishDate | 2022-11-01 |
publisher | Frontiers Media S.A. |
record_format | Article |
series | Frontiers in Cell and Developmental Biology |
spelling | doaj.art-dc9b36bedb3f46beb7c5ce9856ed3ac62022-12-22T04:36:43ZengFrontiers Media S.A.Frontiers in Cell and Developmental Biology2296-634X2022-11-011010.3389/fcell.2022.10507691050769MCI-frcnn: A deep learning method for topological micro-domain boundary detectionSimon Zhongyuan Tian0Pengfei Yin1Kai Jing2Yang Yang3Yewen Xu4Guangyu Huang5Duo Ning6Melissa J. Fullwood7Melissa J. Fullwood8Melissa J. Fullwood9Meizhen Zheng10Shenzhen Key Laboratory of Gene Regulation and Systems Biology, School of Life Sciences, Southern University of Science and Technology, Shenzhen, ChinaShenzhen Key Laboratory of Gene Regulation and Systems Biology, School of Life Sciences, Southern University of Science and Technology, Shenzhen, ChinaShenzhen Key Laboratory of Gene Regulation and Systems Biology, School of Life Sciences, Southern University of Science and Technology, Shenzhen, ChinaShenzhen Key Laboratory of Gene Regulation and Systems Biology, School of Life Sciences, Southern University of Science and Technology, Shenzhen, ChinaShenzhen Key Laboratory of Gene Regulation and Systems Biology, School of Life Sciences, Southern University of Science and Technology, Shenzhen, ChinaShenzhen Key Laboratory of Gene Regulation and Systems Biology, School of Life Sciences, Southern University of Science and Technology, Shenzhen, ChinaShenzhen Key Laboratory of Gene Regulation and Systems Biology, School of Life Sciences, Southern University of Science and Technology, Shenzhen, ChinaSchool of Biological Sciences, Nanyang Technological University, Singapore, SingaporeCancer Science Institute of Singapore, National University of Singapore, Singapore, SingaporeInstitute of Molecular and Cell Biology, Agency for Science, Technology and Research (A*STAR), Singapore, SingaporeShenzhen Key Laboratory of Gene Regulation and Systems Biology, School of Life Sciences, Southern University of Science and Technology, Shenzhen, ChinaChromatin structural domains, or topologically associated domains (TADs), are a general organizing principle in chromatin biology. RNA polymerase II (RNAPII) mediates multiple chromatin interactive loops, tethering together as RNAPII-associated chromatin interaction domains (RAIDs) to offer a framework for gene regulation. RAID and TAD alterations have been found to be associated with diseases. They can be further dissected as micro-domains (micro-TADs and micro-RAIDs) by clustering single-molecule chromatin-interactive complexes from next-generation three-dimensional (3D) genome techniques, such as ChIA-Drop. Currently, there are few tools available for micro-domain boundary identification. In this work, we developed the MCI-frcnn deep learning method to train a Faster Region-based Convolutional Neural Network (Faster R-CNN) for micro-domain boundary detection. At the training phase in MCI-frcnn, 50 images of RAIDs from Drosophila RNAPII ChIA-Drop data, containing 261 micro-RAIDs with ground truth boundaries, were trained for 7 days. Using this well-trained MCI-frcnn, we detected micro-RAID boundaries for the input new images, with a fast speed (5.26 fps), high recognition accuracy (AUROC = 0.85, mAP = 0.69), and high boundary region quantification (genomic IoU = 76%). We further applied MCI-frcnn to detect human micro-TADs boundaries using human GM12878 SPRITE data and obtained a high region quantification score (mean gIoU = 85%). In all, the MCI-frcnn deep learning method which we developed in this work is a general tool for micro-domain boundary detection.https://www.frontiersin.org/articles/10.3389/fcell.2022.1050769/fulldeep learningtopological micro-domainfaster R-CNN algorithm3D genome organizationdomain boundary |
spellingShingle | Simon Zhongyuan Tian Pengfei Yin Kai Jing Yang Yang Yewen Xu Guangyu Huang Duo Ning Melissa J. Fullwood Melissa J. Fullwood Melissa J. Fullwood Meizhen Zheng MCI-frcnn: A deep learning method for topological micro-domain boundary detection Frontiers in Cell and Developmental Biology deep learning topological micro-domain faster R-CNN algorithm 3D genome organization domain boundary |
title | MCI-frcnn: A deep learning method for topological micro-domain boundary detection |
title_full | MCI-frcnn: A deep learning method for topological micro-domain boundary detection |
title_fullStr | MCI-frcnn: A deep learning method for topological micro-domain boundary detection |
title_full_unstemmed | MCI-frcnn: A deep learning method for topological micro-domain boundary detection |
title_short | MCI-frcnn: A deep learning method for topological micro-domain boundary detection |
title_sort | mci frcnn a deep learning method for topological micro domain boundary detection |
topic | deep learning topological micro-domain faster R-CNN algorithm 3D genome organization domain boundary |
url | https://www.frontiersin.org/articles/10.3389/fcell.2022.1050769/full |
work_keys_str_mv | AT simonzhongyuantian mcifrcnnadeeplearningmethodfortopologicalmicrodomainboundarydetection AT pengfeiyin mcifrcnnadeeplearningmethodfortopologicalmicrodomainboundarydetection AT kaijing mcifrcnnadeeplearningmethodfortopologicalmicrodomainboundarydetection AT yangyang mcifrcnnadeeplearningmethodfortopologicalmicrodomainboundarydetection AT yewenxu mcifrcnnadeeplearningmethodfortopologicalmicrodomainboundarydetection AT guangyuhuang mcifrcnnadeeplearningmethodfortopologicalmicrodomainboundarydetection AT duoning mcifrcnnadeeplearningmethodfortopologicalmicrodomainboundarydetection AT melissajfullwood mcifrcnnadeeplearningmethodfortopologicalmicrodomainboundarydetection AT melissajfullwood mcifrcnnadeeplearningmethodfortopologicalmicrodomainboundarydetection AT melissajfullwood mcifrcnnadeeplearningmethodfortopologicalmicrodomainboundarydetection AT meizhenzheng mcifrcnnadeeplearningmethodfortopologicalmicrodomainboundarydetection |