Deep latent space fusion for adaptive representation of heterogeneous multi-omics data

Abstract

The integration of multi-omics data makes it possible to understand complex biological organisms at the system level. Numerous integration approaches have been developed by assuming a common underlying data space. Due to the noise and heterogeneity of biological data, the performance of these approaches is greatly affected. In this work, we propose a novel deep neural network architecture, named Deep Latent Space Fusion (DLSF), which integrates the multi-omics data by learning consistent manifold in the sample latent space for disease subtypes identification. DLSF is built upon a cycle autoencoder with a shared self-expressive layer, which can naturally and adaptively merge nonlinear features at each omics level into one unified sample manifold and produce adaptive representation of heterogeneous samples at the multi-omics level. We have assessed DLSF on various biological and biomedical datasets to validate its effectiveness. DLSF can efficiently and accurately capture the intrinsic manifold of the sample structures or sample clusters compared with other state-of-the-art methods, and DLSF yielded more significant outcomes for biological significance, survival prognosis and clinical relevance in application of cancer study in The Cancer Genome Atlas. Notably, as a deep case study, we determined a new molecular subtype of kidney renal clear cell carcinoma that may benefit immunotherapy in the viewpoint of multi-omics, and we further found potential subtype-specific biomarkers from multiple omics data, which were validated by independent datasets. In addition, we applied DLSF to identify potential therapeutic agents of different molecular subtypes of chronic lymphocytic leukemia, demonstrating the scalability of DLSF in diverse omics data types and application scenarios.

MultiGATAE: A Novel Cancer Subtype Identification Method Based on Multi-Omics and Attention Mechanism

“…To evaluate the performance of our proposed-method multiGATAE, we compared it with eight state-of-the-art clustering methods, namely, DLSF ( Zhang et al, 2022 ), subtype-WESLR ( Song et al, 2021 ), SNF ( Wang et al, 2014 ), NEMO ( Rappoport and Shamir, 2019 ), iClusterBayes ( Mo et al, 2018 ), moCluster ( Meng et al, 2016 ), LRAcluster ( Wu et al, 2015 ), and PFA ( Shi et al, 2017 ) on eight public cancer multi-omics datasets. Here, we first introduce the details of these eight state-of-the-art methods, then we introduce the datasets used in this section and show the experiment results on these eight datasets.…”

Multi-omics of extracellular vesicles: An integrative representation of functional mediators and perspectives on lung disease study

“…Therefore, integrative bioinformatic models and methods ( Yu and Zeng, 2018 ; Zhang et al, 2022 ; Tang et al, 2022 ; Yu et al, 2022 ) should be suitable for inferring the integrative representation of EVs at the molecular and cellular levels, and this in turn can aid in understanding the intracellular journeys of EVs. As it is not limited to the typical integration of genetic information for examining the potential biological pathways in a cell, the new integrative bioinformatic analysis of EVs can bridge the underlying regulatory signal flow among cells by combining with other cutting-edge biotechnology methods such as single-cell omics using a form of vertical integration.…”

MCluster-VAEs: An end-to-end variational deep learning-based clustering method for subtype discovery using multi-omics data

“…Note that cancer subtypes that are biologically different may have similar survival, and this is also true for enrichment of clinical parameters. However, these measures are widely used for clustering assessment, including in the most multi-omics clustering studies (12,26,(36)(37)(38).…”