Improvement of variables interpretability in kernel PCA
Abstract Background Kernel methods have been proven to be a powerful tool for the integration and analysis of high-throughput technologies generated data. Kernels offer a nonlinear version of any linear algorithm solely based on dot products. The kernelized version of principal component analysis is...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
BMC
2023-07-01
|
| Series: | BMC Bioinformatics |
| Subjects: | |
| Online Access: | https://doi.org/10.1186/s12859-023-05404-y |
| _version_ | 1797778617595330560 |
|---|---|
| author | Mitja Briscik Marie-Agnès Dillies Sébastien Déjean |
| author_facet | Mitja Briscik Marie-Agnès Dillies Sébastien Déjean |
| author_sort | Mitja Briscik |
| collection | DOAJ |
| description | Abstract Background Kernel methods have been proven to be a powerful tool for the integration and analysis of high-throughput technologies generated data. Kernels offer a nonlinear version of any linear algorithm solely based on dot products. The kernelized version of principal component analysis is a valid nonlinear alternative to tackle the nonlinearity of biological sample spaces. This paper proposes a novel methodology to obtain a data-driven feature importance based on the kernel PCA representation of the data. Results The proposed method, kernel PCA Interpretable Gradient (KPCA-IG), provides a data-driven feature importance that is computationally fast and based solely on linear algebra calculations. It has been compared with existing methods on three benchmark datasets. The accuracy obtained using KPCA-IG selected features is equal to or greater than the other methods’ average. Also, the computational complexity required demonstrates the high efficiency of the method. An exhaustive literature search has been conducted on the selected genes from a publicly available Hepatocellular carcinoma dataset to validate the retained features from a biological point of view. The results once again remark on the appropriateness of the computed ranking. Conclusions The black-box nature of kernel PCA needs new methods to interpret the original features. Our proposed methodology KPCA-IG proved to be a valid alternative to select influential variables in high-dimensional high-throughput datasets, potentially unravelling new biological and medical biomarkers. |
| first_indexed | 2024-03-12T23:20:00Z |
| format | Article |
| id | doaj.art-18fb56440aad4494ac5a8ec171515093 |
| institution | Directory Open Access Journal |
| issn | 1471-2105 |
| language | English |
| last_indexed | 2024-03-12T23:20:00Z |
| publishDate | 2023-07-01 |
| publisher | BMC |
| record_format | Article |
| series | BMC Bioinformatics |
| spelling | doaj.art-18fb56440aad4494ac5a8ec1715150932023-07-16T11:29:53ZengBMCBMC Bioinformatics1471-21052023-07-0124112110.1186/s12859-023-05404-yImprovement of variables interpretability in kernel PCAMitja Briscik0Marie-Agnès Dillies1Sébastien Déjean2Institut de Mathématiques de Toulouse, UMR5219, CNRS, UPS, Université de ToulouseInstitut Pasteur, Université Paris Cité, Bioinformatics and Biostatistics HubInstitut de Mathématiques de Toulouse, UMR5219, CNRS, UPS, Université de ToulouseAbstract Background Kernel methods have been proven to be a powerful tool for the integration and analysis of high-throughput technologies generated data. Kernels offer a nonlinear version of any linear algorithm solely based on dot products. The kernelized version of principal component analysis is a valid nonlinear alternative to tackle the nonlinearity of biological sample spaces. This paper proposes a novel methodology to obtain a data-driven feature importance based on the kernel PCA representation of the data. Results The proposed method, kernel PCA Interpretable Gradient (KPCA-IG), provides a data-driven feature importance that is computationally fast and based solely on linear algebra calculations. It has been compared with existing methods on three benchmark datasets. The accuracy obtained using KPCA-IG selected features is equal to or greater than the other methods’ average. Also, the computational complexity required demonstrates the high efficiency of the method. An exhaustive literature search has been conducted on the selected genes from a publicly available Hepatocellular carcinoma dataset to validate the retained features from a biological point of view. The results once again remark on the appropriateness of the computed ranking. Conclusions The black-box nature of kernel PCA needs new methods to interpret the original features. Our proposed methodology KPCA-IG proved to be a valid alternative to select influential variables in high-dimensional high-throughput datasets, potentially unravelling new biological and medical biomarkers.https://doi.org/10.1186/s12859-023-05404-yKernel PCARelevant variablesUnsupervised learningKernel methods |
| spellingShingle | Mitja Briscik Marie-Agnès Dillies Sébastien Déjean Improvement of variables interpretability in kernel PCA BMC Bioinformatics Kernel PCA Relevant variables Unsupervised learning Kernel methods |
| title | Improvement of variables interpretability in kernel PCA |
| title_full | Improvement of variables interpretability in kernel PCA |
| title_fullStr | Improvement of variables interpretability in kernel PCA |
| title_full_unstemmed | Improvement of variables interpretability in kernel PCA |
| title_short | Improvement of variables interpretability in kernel PCA |
| title_sort | improvement of variables interpretability in kernel pca |
| topic | Kernel PCA Relevant variables Unsupervised learning Kernel methods |
| url | https://doi.org/10.1186/s12859-023-05404-y |
| work_keys_str_mv | AT mitjabriscik improvementofvariablesinterpretabilityinkernelpca AT marieagnesdillies improvementofvariablesinterpretabilityinkernelpca AT sebastiendejean improvementofvariablesinterpretabilityinkernelpca |