Structural Learning and Integrative Decomposition of Multi-View Data

Gaynanova, Irina; Li, Gen

doi:10.1111/biom.13108

Cited by 67 publications

(72 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the context of vertical integration, the joint and individual scores

boldV

and

V_{i}

have been applied to risk prediction (Kaplan and Lock, ) and clustering (Hellton and Thoresen, ) for high‐dimensional data. Several related techniques, such as AJIVE (Feng et al ., ) and SLIDE (Gaynanova and Li, ), have been proposed (Zhou et al ., ), as well as extensions that allow the adjustment of covariates (Li and Jung, ) or accommodate heterogeneity in the distributional assumptions for different sources (Li and Gaynanova, ; Zhu et al ., ).…”

Section: Introductionmentioning

confidence: 99%

Integrative factorization of bidimensionally linked matrices

Park

Lock

2019

Biometrics

View full text Add to dashboard Cite

Advances in molecular “omics” technologies have motivated new methodologies for the integration of multiple sources of high‐content biomedical data. However, most statistical methods for integrating multiple data matrices only consider data shared vertically (one cohort on multiple platforms) or horizontally (different cohorts on a single platform). This is limiting for data that take the form of bidimensionally linked matrices (eg, multiple cohorts measured on multiple platforms), which are increasingly common in large‐scale biomedical studies. In this paper, we propose bidimensional integrative factorization (BIDIFAC) for integrative dimension reduction and signal approximation of bidimensionally linked data matrices. Our method factorizes data into (a) globally shared, (b) row‐shared, (c) column‐shared, and (d) single‐matrix structural components, facilitating the investigation of shared and unique patterns of variability. For estimation, we use a penalized objective function that extends the nuclear norm penalization for a single matrix. As an alternative to the complicated rank selection problem, we use results from the random matrix theory to choose tuning parameters. We apply our method to integrate two genomics platforms (messenger RNA and microRNA expression) across two sample cohorts (tumor samples and normal tissue samples) using the breast cancer data from the Cancer Genome Atlas. We provide R code for fitting BIDIFAC, imputing missing values, and generating simulated data.

show abstract

“…In the context of vertical integration, the joint and individual scores

boldV

and

V_{i}

Section: Introductionmentioning

confidence: 99%

Integrative factorization of bidimensionally linked matrices

Park

Lock

2019

Biometrics

View full text Add to dashboard Cite

show abstract

“…This is a limitation of most data integration methods, and we expect partially shared components to result in even better prediction models. A way to investigate partially shared patterns is provided in the SLIDE method by [37], and is a potential starting point for further work in this direction.…”

Section: Discussionmentioning

confidence: 99%

Integrative Analysis of Multi-omics Data Improves Model Predictions: An Application to Lung Cancer

Ponzi

Thoresen

Nøst

et al. 2020

Preprint

View full text Add to dashboard Cite

Background: Cancer genomic studies often include data collected from several omics platforms. Each omics data source contributes to the understanding of the underlying biological process via source speciﬁc (”individual”) patterns of variability. At the same time, statistical associations and potential interactions among the diﬀerent data sources can reveal signals from common biological processes that might not be identiﬁed by single source analyses. These common patterns of variability are referred to as ”shared” or ”joint”. To capture both contributions of variance, integrative dimension reduction techniques are needed. Integrated PCA is a model based generalization of principal components analysis that separates shared and source speciﬁc variance by iteratively estimating covariance structures from a matrix normal distribution. Angle based JIVE is a matrix factorization method that decomposes joint and individual variation by permutation of row subspaces. We apply these techniques to identify joint and individual contributions of DNA methylation, miRNA and mRNA expression collected from blood samples in a lung cancer case control study nested within the Norwegian Woman and Cancer (NOWAC) cohort study.Results: In this work, we show how an integrative analysis that preserves both components of variation is more appropriate than analyses considering uniquely individual or joint components. Our results show how both joint and individual components contribute to a better quality of model predictions, and facilitate the interpretation of the underlying biological processes.Conclusions: When compared to a non integrative analysis of the three omics sources, integrative models that simultaneously include joint and individual components result in better prediction of cancer status and metastatic cancer at diagnosis.

show abstract

“…This is critically important for generating meaningful, interpretable results using existing statistical models (45,46). Other methods, such as structural learning and integrative decomposition of multi-view data (47), common orthogonal basis extraction (48), and group factor analysis (49), may also be used. Our results suggest that an attractive feature of JIVE is the performance robustness, consistent with our prior study of brain age prediction (20).…”

Section: Structural Covariation May Reflect Synchronized Developmentmentioning

confidence: 99%

Brain Anatomical Covariation Patterns Linked to Binge Drinking and Age at First Full Drink Prior to 21 Years

Zhao

Constable

Hien

et al. 2020

Preprint

View full text Add to dashboard Cite

Binge drinking and age at first full drink of alcohol prior to 21 years (AFD<21) have been linked to neuroanatomical differences in cortical and subcortical grey matter (GM) volume, cortical thickness, and surface area. Despite the potential to reveal novel network-level relationships, structural covariation patterns among these morphological measures have yet to be examined relative to binge drinking and AFD<21. Here, we used the Joint and Individual Variance Explained (JIVE) method to characterize structural covariation patterns common across and specific to morphological measures in 293 participants (149 individuals with binge drinking and 144 healthy controls) from the Human Connectome Project (HCP). An independent dataset (Nathan Kline Institute Rockland Sample; NKI-RS) was used to examine reproducibility/ generalizability. We identified a highly reproducible joint component dominated by structural covariation between GM volume in the brainstem and thalamus proper, and GM volume and surface area in prefrontal cortical regions. Using linear mixed regression models, we found that this joint component was related to AFD<21 in both the HCP and NKI-RS datasets, whereas the individual thickness component associated with binge drinking and AFD<21 in the HCP dataset was not statistically significant in the NKI-RS sample. Taken together, our results show that a highly reproducible structural pattern involving covariation in brain regions relevant to thalamic-PFC-brainstem neural circuitry is linked to age at first full drink.

show abstract

Structural Learning and Integrative Decomposition of Multi-View Data

Cited by 67 publications

References 27 publications

Integrative factorization of bidimensionally linked matrices

Integrative factorization of bidimensionally linked matrices

Integrative Analysis of Multi-omics Data Improves Model Predictions: An Application to Lung Cancer

Brain Anatomical Covariation Patterns Linked to Binge Drinking and Age at First Full Drink Prior to 21 Years

Contact Info

Product

Resources

About