Integrative and regularized principal component analysis of multiple sources of data

Abstract

Integration of data of disparate types has become increasingly important to enhancing the power for new discoveries by combining complementary strengths of multiple types of data. One application is to uncover tumor subtypes in human cancer research, in which multiple types of genomic data are integrated, including gene expression, DNA copy number and DNA methylation data. In spite of their successes, existing approaches based on joint latent variable models require stringent distributional assumptions and may suffer from unbalanced scales (or units) of different types of data and non-scalability of the corresponding algorithms. In this paper, we propose an alternative based on integrative and regularized principal component analysis, which is distribution-free, computationally efficient, and robust against unbalanced scales. The new method performs dimension reduction simultaneously on multiple types of data, seeking data-adaptive sparsity and scaling. As a result, in addition to feature selection for each type of data, integrative clustering is achieved. Numerically, the proposed method compares favorably against its competitors in terms of accuracy (in identifying hidden clusters), computational efficiency, and robustness against unbalanced scales. In particular, compared to a popular method, the new method was competitive in identifying tumor subtypes associated with distinct patient survival patterns when applied to a combined analysis of DNA copy number, mRNA expression and DNA methylation data in a glioblastoma multiforme study.

Validation of a predictive method for sunscreen formula evaluation using gene expression analysis in a Chinese reconstructed full‐thickness skin model

“…For DUVR‐induced gene expression modulation and secreted protein analysis the means were compared using two‐tailed Student's t ‐test, difference was considered significant at P < 0.05 . For sunscreen ranking, principal component analysis (PCA) was conducted by normalizing the modulation ratio of all the genes to a notionally common scale and analysed by the principal of dimension reduction as previously described . Dimension reduction (factor analysis) was run with SPSS, Chicago, IL, U.S.A., components were extracted based on eigenvalues ( λ 1 , λ 2 , … λ n ) >1 and factor score of each component was got ( F 1 , F 2 , … F n ).…”

Evaluation and comparison of multi-omics data integration methods for cancer subtyping

“…To understand and demonstrate the crucial need for addressing the second problem of data type selection, we surveyed 58 integration methods for cancer subtyping proposed from 2009 to 2019, and the result is summarized in Fig 1 where gene expression is treated as the same as mRNA expression and miRNA expression is placed into the group of epigenome based on observations from [7]. We summarized part of these 58 integration methods with the omics data they used in Fig 1A , and we can see, the data combinations used in these methods [2,[8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23][24] are significantly inconsistent. For example, Fig 1B shows that while the mRNA expression data were used by 56 of the 58 methods, each of the other data types was only used by at most nearly half of these methods.…”

Integrative factorization of bidimensionally linked matrices

“…To capture joint variation, concatenated PCA assumes for each matrix , that is, the scores are shared across matrices. The iCluster (Shen et al ., ) and irPCA (Liu et al ., ) approaches make this assumption for the integration of multisource biomedical data. Alternatively, more flexible approaches allow for structured variation that may be shared across matrices or specific to individual matrices.…”