An official website of the United States government
Abstract MotivationIntegrating multiple omics datasets can significantly advance our understanding of disease mechanisms, physiology, and treatment responses. However, a major challenge in multi-omics studies is the disparity in sample sizes across different datasets, which can introduce bias and reduce statistical power. To address this issue, we propose a novel framework, OmicsNMF, designed to impute missing omics data and enhance disease phenotype prediction. OmicsNMF integrates Generative Adversarial Networks (GANs) with Non-Negative Matrix Factorization (NMF). NMF is a well-established method for uncovering underlying patterns in omics data, while GANs enhance the imputation process by generating realistic data samples. This synergy aims to more effectively address sample size disparity, thereby improving data integration and prediction accuracy. ResultsFor evaluation, we focused on predicting breast cancer subtypes using the imputed data generated by our proposed framework, OmicsNMF. Our results indicate that OmicsNMF consistently outperforms baseline methods. We further assessed the quality of the imputed data through survival analysis, revealing that the imputed omics profiles provide significant prognostic power for both overall survival and disease-free status. Overall, OmicsNMF effectively leverages GANs and NMF to impute missing samples while preserving key biological features. This approach shows potential for advancing precision oncology by improving data integration and analysis. Availability and implementationSource code is available at: https://github.com/compbiolabucf/OmicsNMF.
more »
« less
Acute stress reduces population-level metabolic and proteomic variation
Abstract BackgroundVariation in omics data due to intrinsic biological stochasticity is often viewed as a challenging and undesirable feature of complex systems analyses. In fact, numerous statistical methods are utilized to minimize the variation among biological replicates. ResultsWe demonstrate that the common statistics relative standard deviation (RSD) and coefficient of variation (CV), which are often used for quality control or part of a larger pipeline in omics analyses, can also be used as a metric of a physiological stress response. Using an approach we term Replicate Variation Analysis (RVA), we demonstrate that acute physiological stress leads to feature-wide canalization of CV profiles of metabolomes and proteomes across biological replicates. Canalization is the repression of variation between replicates, which increases phenotypic similarity. Multiple in-house mass spectrometry omics datasets in addition to publicly available data were analyzed to assess changes in CV profiles in plants, animals, and microorganisms. In addition, proteomics data sets were evaluated utilizing RVA to identify functionality of reduced CV proteins. ConclusionsRVA provides a foundation for understanding omics level shifts that occur in response to cellular stress. This approach to data analysis helps characterize stress response and recovery, and could be deployed to detect populations under stress, monitor health status, and conduct environmental monitoring.
more »
« less
- Award ID(s):
- 1929113
- PAR ID:
- 10491369
- Publisher / Repository:
- BMC Springer Nature
- Date Published:
- Journal Name:
- BMC Bioinformatics
- Volume:
- 24
- Issue:
- 1
- ISSN:
- 1471-2105
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Abstract MotivationPredictive biological signatures provide utility as biomarkers for disease diagnosis and prognosis, as well as prediction of responses to vaccination or therapy. These signatures are identified from high-throughput profiling assays through a combination of dimensionality reduction and machine learning techniques. The genes, proteins, metabolites, and other biological analytes that compose signatures also generate hypotheses on the underlying mechanisms driving biological responses, thus improving biological understanding. Dimensionality reduction is a critical step in signature discovery to address the large number of analytes in omics datasets, especially for multi-omics profiling studies with tens of thousands of measurements. Latent factor models, which can account for the structural heterogeneity across diverse assays, effectively integrate multi-omics data and reduce dimensionality to a small number of factors that capture correlations and associations among measurements. These factors provide biologically interpretable features for predictive modeling. However, multi-omics integration and predictive modeling are generally performed independently in sequential steps, leading to suboptimal factor construction. Combining these steps can yield better multi-omics signatures that are more predictive while still being biologically meaningful. ResultsWe developed a supervised variational Bayesian factor model that extracts multi-omics signatures from high-throughput profiling datasets that can span multiple data types. Signature-based multiPle-omics intEgration via lAtent factoRs (SPEAR) adaptively determines factor rank, emphasis on factor structure, data relevance and feature sparsity. The method improves the reconstruction of underlying factors in synthetic examples and prediction accuracy of coronavirus disease 2019 severity and breast cancer tumor subtypes. Availability and implementationSPEAR is a publicly available R-package hosted at https://bitbucket.org/kleinstein/SPEAR.more » « less
-
Abstract BackgroundThe recent development of high-throughput sequencing has created a large collection of multi-omics data, which enables researchers to better investigate cancer molecular profiles and cancer taxonomy based on molecular subtypes. Integrating multi-omics data has been proven to be effective for building more precise classification models. Most current multi-omics integrative models use either an early fusion in the form of concatenation or late fusion with a separate feature extractor for each omic, which are mainly based on deep neural networks. Due to the nature of biological systems, graphs are a better structural representation of bio-medical data. Although few graph neural network (GNN) based multi-omics integrative methods have been proposed, they suffer from three common disadvantages. One is most of them use only one type of connection, either inter-omics or intra-omic connection; second, they only consider one kind of GNN layer, either graph convolution network (GCN) or graph attention network (GAT); and third, most of these methods have not been tested on a more complex classification task, such as cancer molecular subtypes. ResultsIn this study, we propose a novel end-to-end multi-omics GNN framework for accurate and robust cancer subtype classification. The proposed model utilizes multi-omics data in the form of heterogeneous multi-layer graphs, which combine both inter-omics and intra-omic connections from established biological knowledge. The proposed model incorporates learned graph features and global genome features for accurate classification. We tested the proposed model on the Cancer Genome Atlas (TCGA) Pan-cancer dataset and TCGA breast invasive carcinoma (BRCA) dataset for molecular subtype and cancer subtype classification, respectively. The proposed model shows superior performance compared to four current state-of-the-art baseline models in terms of accuracy, F1 score, precision, and recall. The comparative analysis of GAT-based models and GCN-based models reveals that GAT-based models are preferred for smaller graphs with less information and GCN-based models are preferred for larger graphs with extra information.more » « less
-
ABSTRACT ObjectivesThe impacts of stress on inflammation, although hypothesized, have not been thoroughly examined, especially in relation to social and environmental factors and particularly within Black populations. This study aims to explore the biological mechanisms of embodiment linking stress and health to understand physiological changes in the body's response to psychological stress in a Nigerian population. Through a multidisciplinary approach, this study queries the relationship between stress, cortisol, and salivary C‐reactive protein (sCRP), a biomarker of inflammation, while also validating the use of sCRP as a potential and accurate stress indicator in the field. MethodsIn this cross‐sectional study, 138 passive drool saliva samples (nfemale = 89nmale = 49) were collected and assessed for sCRP and cortisol levels in adults. Participants also completed a short demographic survey and, to measure psychological stress, the General Health Questionnaire 12 (GHQ‐12). Relationships between sCRP and stress‐related variables (i.e., cortisol, GHQ‐12, and demographic data) were assessed using Spearman's correlations, simple regression, multivariable linear regression, and exploratory factor analysis. ResultssCRP levels ranged from 20.57 to 6879.41 pg/mL across all samples, with significant differences between female and male participants. The GHQ‐12 was not a significant predictor of sCRP variability. However, socio‐demographic factors such as body mass index (BMI), age, self‐reported sex, ethnic identity, and cortisol were significant predictors, collectively explaining 24%–27% of the variation in sCRP. ConclusionSocio‐demographic predictors like BMI, age, sex, and particularly ethnic group experience in Nigeria encapsulate aspects of embodied stress, that significantly affect sCRP variability.more » « less
-
Abstract BackgroundPrior work described an approach for mapping the two-dimensional spatial distribution of biaxial residual stress in plate-like samples, the approach combining multiple slitting measurements with elastic stress analysis. Objective This paper extends the prior work by applying a new variation of the slitting method that uses measurements of cut mouth opening displacement (CMOD) rather than back-face strain (BFS). MethodsFirst, CMOD slitting is validated using an experiment where: BFS and CMOD are measured simultaneously on the same sample during incremental slitting; two residual stress profiles are computed, one from the BFS data and a second from the CMOD data; and the two residual stress profiles are compared. Following validation, multiple adjacent CMOD slitting measurements are used to construct two-dimensional maps of residual stress in plates cut from quenched aluminum. ResultsThe two residual stress versus depth profiles, each computed separately from BFS or CMOD data, are in agreement, with compression near the plate boundaries (-150 MPa) and tension near the plate center (100 MPa); differences between the two stress profiles have a maximum of 25 MPa and a RMS of 7.2 MPa. Repeated biaxial residual stress mapping measurements show the CMOD technique is repeatable, and complementary contour method measurements show the mappings are valid. Aspects of CMOD and BFS deformations during slitting are also described and show they are generally complementary but that CMOD slitting is favorable in narrow samples.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Journal Article:
- https://doi.org/10.1186/s12859-023-05185-4
