Autoencoders with shared and specific embeddings for multi-omics data integration

Abstract Background In cancer research, different levels of high-dimensional data are often collected for the same subjects. Effective integration of these data by considering the shared and specific information from each data source can help us better understand different types of cancer. Results I...

Full description

Saved in:
Bibliographic Details
Main Authors: Chao Wang, Michael J. O’Connell
Format: Article
Language:English
Published: BMC 2025-08-01
Series:BMC Bioinformatics
Subjects:
Online Access:https://doi.org/10.1186/s12859-025-06245-7
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849225849293766656
author Chao Wang
Michael J. O’Connell
author_facet Chao Wang
Michael J. O’Connell
author_sort Chao Wang
collection DOAJ
description Abstract Background In cancer research, different levels of high-dimensional data are often collected for the same subjects. Effective integration of these data by considering the shared and specific information from each data source can help us better understand different types of cancer. Results In this study we propose a novel autoencoder (AE) structure with explicitly defined orthogonal loss between the shared and specific embeddings to integrate different data sources. We compare our model with previously proposed AE structures based on simulated data and real cancer data from The Cancer Genome Atlas. Using simulations with different proportions of differentially expressed genes, we compare the performance of AE methods for subsequent classification tasks. We also compare the model performance with a commonly used dimension reduction method, joint and individual variance explained (JIVE). In terms of reconstruction loss, our proposed AE models with orthogonal constraints have a slightly better reconstruction loss. All AE models achieve higher classification accuracy than the original features, demonstrating the usefulness of the embeddings extracted by the model. Conclusions We show that the proposed models have consistently high classification accuracy on both training and testing sets. In comparison, the recently proposed MOCSS model that imposes an orthogonality penalty in the post-processing step has lower classification accuracy that is on par with JIVE.
format Article
id doaj-art-84401cc432874bd9a04cf0e1f7bf2999
institution Kabale University
issn 1471-2105
language English
publishDate 2025-08-01
publisher BMC
record_format Article
series BMC Bioinformatics
spelling doaj-art-84401cc432874bd9a04cf0e1f7bf29992025-08-24T11:54:37ZengBMCBMC Bioinformatics1471-21052025-08-0126111610.1186/s12859-025-06245-7Autoencoders with shared and specific embeddings for multi-omics data integrationChao Wang0Michael J. O’Connell1Ben May Department for Cancer Research, University of ChicagoDepartment of Statistics, Miami UniversityAbstract Background In cancer research, different levels of high-dimensional data are often collected for the same subjects. Effective integration of these data by considering the shared and specific information from each data source can help us better understand different types of cancer. Results In this study we propose a novel autoencoder (AE) structure with explicitly defined orthogonal loss between the shared and specific embeddings to integrate different data sources. We compare our model with previously proposed AE structures based on simulated data and real cancer data from The Cancer Genome Atlas. Using simulations with different proportions of differentially expressed genes, we compare the performance of AE methods for subsequent classification tasks. We also compare the model performance with a commonly used dimension reduction method, joint and individual variance explained (JIVE). In terms of reconstruction loss, our proposed AE models with orthogonal constraints have a slightly better reconstruction loss. All AE models achieve higher classification accuracy than the original features, demonstrating the usefulness of the embeddings extracted by the model. Conclusions We show that the proposed models have consistently high classification accuracy on both training and testing sets. In comparison, the recently proposed MOCSS model that imposes an orthogonality penalty in the post-processing step has lower classification accuracy that is on par with JIVE.https://doi.org/10.1186/s12859-025-06245-7Multi-omicsData integrationAutoencodersShared and specific information
spellingShingle Chao Wang
Michael J. O’Connell
Autoencoders with shared and specific embeddings for multi-omics data integration
BMC Bioinformatics
Multi-omics
Data integration
Autoencoders
Shared and specific information
title Autoencoders with shared and specific embeddings for multi-omics data integration
title_full Autoencoders with shared and specific embeddings for multi-omics data integration
title_fullStr Autoencoders with shared and specific embeddings for multi-omics data integration
title_full_unstemmed Autoencoders with shared and specific embeddings for multi-omics data integration
title_short Autoencoders with shared and specific embeddings for multi-omics data integration
title_sort autoencoders with shared and specific embeddings for multi omics data integration
topic Multi-omics
Data integration
Autoencoders
Shared and specific information
url https://doi.org/10.1186/s12859-025-06245-7
work_keys_str_mv AT chaowang autoencoderswithsharedandspecificembeddingsformultiomicsdataintegration
AT michaeljoconnell autoencoderswithsharedandspecificembeddingsformultiomicsdataintegration