MultiCapsNet: A General Framework for Data Integration and Interpretable Classification

The latest progresses of experimental biology have generated a large number of data with different formats and lengths. Deep learning is an ideal tool to deal with complex datasets, but its inherent “black box” nature needs more interpretability. At the same time, traditional interpretable machine l...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Lifei Wang, Xuexia Miao, Rui Nie, Zhang Zhang, Jiang Zhang, Jun Cai
Formato:	article
Lenguaje:	EN
Publicado:	Frontiers Media S.A. 2021
Materias:	capsule network classification data integration interpretability modular feature Genetics QH426-470
Acceso en línea:	https://doaj.org/article/5accd7e0eded434394e3092cb1cfcb17
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:5accd7e0eded434394e3092cb1cfcb17
record_format	dspace
spelling	oai:doaj.org-article:5accd7e0eded434394e3092cb1cfcb172021-11-30T19:06:14ZMultiCapsNet: A General Framework for Data Integration and Interpretable Classification1664-802110.3389/fgene.2021.767602https://doaj.org/article/5accd7e0eded434394e3092cb1cfcb172021-11-01T00:00:00Zhttps://www.frontiersin.org/articles/10.3389/fgene.2021.767602/fullhttps://doaj.org/toc/1664-8021The latest progresses of experimental biology have generated a large number of data with different formats and lengths. Deep learning is an ideal tool to deal with complex datasets, but its inherent “black box” nature needs more interpretability. At the same time, traditional interpretable machine learning methods, such as linear regression or random forest, could only deal with numerical features instead of modular features often encountered in the biological field. Here, we present MultiCapsNet (https://github.com/wanglf19/MultiCapsNet), a new deep learning model built on CapsNet and scCapsNet, which possesses the merits such as easy data integration and high model interpretability. To demonstrate the ability of this model as an interpretable classifier to deal with modular inputs, we test MultiCapsNet on three datasets with different data type and application scenarios. Firstly, on the labeled variant call dataset, MultiCapsNet shows a similar classification performance with neural network model, and provides importance scores for data sources directly without an extra importance determination step required by the neural network model. The importance scores generated by these two models are highly correlated. Secondly, on single cell RNA sequence (scRNA-seq) dataset, MultiCapsNet integrates information about protein-protein interaction (PPI), and protein-DNA interaction (PDI). The classification accuracy of MultiCapsNet is comparable to the neural network and random forest model. Meanwhile, MultiCapsNet reveals how each transcription factor (TF) or PPI cluster node contributes to classification of cell type. Thirdly, we made a comparison between MultiCapsNet and SCENIC. The results show several cell type relevant TFs identified by both methods, further proving the validity and interpretability of the MultiCapsNet.Lifei WangLifei WangLifei WangLifei WangXuexia MiaoXuexia MiaoRui NieRui NieRui NieZhang ZhangJiang ZhangJun CaiJun CaiJun CaiFrontiers Media S.A.articlecapsule networkclassificationdata integrationinterpretabilitymodular featureGeneticsQH426-470ENFrontiers in Genetics, Vol 12 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	capsule network classification data integration interpretability modular feature Genetics QH426-470
spellingShingle	capsule network classification data integration interpretability modular feature Genetics QH426-470 Lifei Wang Lifei Wang Lifei Wang Lifei Wang Xuexia Miao Xuexia Miao Rui Nie Rui Nie Rui Nie Zhang Zhang Jiang Zhang Jun Cai Jun Cai Jun Cai MultiCapsNet: A General Framework for Data Integration and Interpretable Classification
description	The latest progresses of experimental biology have generated a large number of data with different formats and lengths. Deep learning is an ideal tool to deal with complex datasets, but its inherent “black box” nature needs more interpretability. At the same time, traditional interpretable machine learning methods, such as linear regression or random forest, could only deal with numerical features instead of modular features often encountered in the biological field. Here, we present MultiCapsNet (https://github.com/wanglf19/MultiCapsNet), a new deep learning model built on CapsNet and scCapsNet, which possesses the merits such as easy data integration and high model interpretability. To demonstrate the ability of this model as an interpretable classifier to deal with modular inputs, we test MultiCapsNet on three datasets with different data type and application scenarios. Firstly, on the labeled variant call dataset, MultiCapsNet shows a similar classification performance with neural network model, and provides importance scores for data sources directly without an extra importance determination step required by the neural network model. The importance scores generated by these two models are highly correlated. Secondly, on single cell RNA sequence (scRNA-seq) dataset, MultiCapsNet integrates information about protein-protein interaction (PPI), and protein-DNA interaction (PDI). The classification accuracy of MultiCapsNet is comparable to the neural network and random forest model. Meanwhile, MultiCapsNet reveals how each transcription factor (TF) or PPI cluster node contributes to classification of cell type. Thirdly, we made a comparison between MultiCapsNet and SCENIC. The results show several cell type relevant TFs identified by both methods, further proving the validity and interpretability of the MultiCapsNet.
format	article
author	Lifei Wang Lifei Wang Lifei Wang Lifei Wang Xuexia Miao Xuexia Miao Rui Nie Rui Nie Rui Nie Zhang Zhang Jiang Zhang Jun Cai Jun Cai Jun Cai
author_facet	Lifei Wang Lifei Wang Lifei Wang Lifei Wang Xuexia Miao Xuexia Miao Rui Nie Rui Nie Rui Nie Zhang Zhang Jiang Zhang Jun Cai Jun Cai Jun Cai
author_sort	Lifei Wang
title	MultiCapsNet: A General Framework for Data Integration and Interpretable Classification
title_short	MultiCapsNet: A General Framework for Data Integration and Interpretable Classification
title_full	MultiCapsNet: A General Framework for Data Integration and Interpretable Classification
title_fullStr	MultiCapsNet: A General Framework for Data Integration and Interpretable Classification
title_full_unstemmed	MultiCapsNet: A General Framework for Data Integration and Interpretable Classification
title_sort	multicapsnet: a general framework for data integration and interpretable classification
publisher	Frontiers Media S.A.
publishDate	2021
url	https://doaj.org/article/5accd7e0eded434394e3092cb1cfcb17
work_keys_str_mv	AT lifeiwang multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT lifeiwang multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT lifeiwang multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT lifeiwang multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT xuexiamiao multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT xuexiamiao multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT ruinie multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT ruinie multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT ruinie multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT zhangzhang multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT jiangzhang multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT juncai multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT juncai multicapsnetageneralframeworkfordataintegrationandinterpretableclassification AT juncai multicapsnetageneralframeworkfordataintegrationandinterpretableclassification
_version_	1718406325024587776

MultiCapsNet: A General Framework for Data Integration and Interpretable Classification

Ejemplares similares