Integrative analysis of multiple diverse omics datasets by sparse group multitask regression
A variety of high throughput genome-wide assays enable the exploration of genetic risk factors underlying complex traits. Although these studies have remarkable impact on identifying susceptible biomarkers, they suffer from issues such as limited sample size and low reproducibility. Combining indivi...
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2014-10-01
|
| Series: | Frontiers in Cell and Developmental Biology |
| Subjects: | |
| Online Access: | http://journal.frontiersin.org/Journal/10.3389/fcell.2014.00062/full |
| _version_ | 1818561934670168064 |
|---|---|
| author | Dongdong eLin Dongdong eLin Jigang eZhang Jigang eZhang Jingyao eLi Jingyao eLi Hao eHe Hao eHe hong-wen eDeng hong-wen eDeng Yu-Ping eWang Yu-Ping eWang Yu-Ping eWang |
| author_facet | Dongdong eLin Dongdong eLin Jigang eZhang Jigang eZhang Jingyao eLi Jingyao eLi Hao eHe Hao eHe hong-wen eDeng hong-wen eDeng Yu-Ping eWang Yu-Ping eWang Yu-Ping eWang |
| author_sort | Dongdong eLin |
| collection | DOAJ |
| description | A variety of high throughput genome-wide assays enable the exploration of genetic risk factors underlying complex traits. Although these studies have remarkable impact on identifying susceptible biomarkers, they suffer from issues such as limited sample size and low reproducibility. Combining individual studies of different genetic levels/platforms has the promise to improve the power and consistency of biomarker identification. In this paper, we propose a novel integrative method, namely sparse group multitask regression, for integrating diverse omics datasets, platforms and populations to identify risk genes/factors of complex diseases. This method combines multitask learning with sparse group regularization, which will: 1) treat the biomarker identification in each single study as a task and then combine them by multitask learning; 2) group variables from all studies for identifying significant genes; 3) enforce sparse constraint on groups of variables to overcome the ‘small sample, but large variables’ problem. We introduce two sparse group penalties: sparse group lasso and sparse group ridge in our multitask model, and provide an effective algorithm for each model. In addition, we propose a significance test for the identification of potential risk genes. Two simulation studies are performed to evaluate the performance of our integrative method by comparing it with conventional meta-analysis method. The results show that our sparse group multitask method outperforms meta-analysis method significantly. In an application to our osteoporosis studies, 7 genes are identified as significant genes by our method and are found to have significant effects in other three independent studies for validation. The most significant gene SOD2 has been identified in our previous osteoporosis study involving the same expression dataset. Several other genes such as TREML2, HTR1E and GLO1 are shown to be novel susceptible genes for osteoporosis, as confirmed from other studies. |
| first_indexed | 2024-12-14T00:57:12Z |
| format | Article |
| id | doaj.art-31082346db13453881fa98555eafb71b |
| institution | Directory Open Access Journal |
| issn | 2296-634X |
| language | English |
| last_indexed | 2024-12-14T00:57:12Z |
| publishDate | 2014-10-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Cell and Developmental Biology |
| spelling | doaj.art-31082346db13453881fa98555eafb71b2022-12-21T23:23:29ZengFrontiers Media S.A.Frontiers in Cell and Developmental Biology2296-634X2014-10-01210.3389/fcell.2014.00062105323Integrative analysis of multiple diverse omics datasets by sparse group multitask regressionDongdong eLin0Dongdong eLin1Jigang eZhang2Jigang eZhang3Jingyao eLi4Jingyao eLi5Hao eHe6Hao eHe7hong-wen eDeng8hong-wen eDeng9Yu-Ping eWang10Yu-Ping eWang11Yu-Ping eWang12Tulane UniversityTulane UniversityTulane UniversityTulane UniversityTulane UniversityTulane UniversityTulane UniversityTulane UniversityTulane UniversityTulane UniversityTulane UniversityTulane UniversityTulane UniversityA variety of high throughput genome-wide assays enable the exploration of genetic risk factors underlying complex traits. Although these studies have remarkable impact on identifying susceptible biomarkers, they suffer from issues such as limited sample size and low reproducibility. Combining individual studies of different genetic levels/platforms has the promise to improve the power and consistency of biomarker identification. In this paper, we propose a novel integrative method, namely sparse group multitask regression, for integrating diverse omics datasets, platforms and populations to identify risk genes/factors of complex diseases. This method combines multitask learning with sparse group regularization, which will: 1) treat the biomarker identification in each single study as a task and then combine them by multitask learning; 2) group variables from all studies for identifying significant genes; 3) enforce sparse constraint on groups of variables to overcome the ‘small sample, but large variables’ problem. We introduce two sparse group penalties: sparse group lasso and sparse group ridge in our multitask model, and provide an effective algorithm for each model. In addition, we propose a significance test for the identification of potential risk genes. Two simulation studies are performed to evaluate the performance of our integrative method by comparing it with conventional meta-analysis method. The results show that our sparse group multitask method outperforms meta-analysis method significantly. In an application to our osteoporosis studies, 7 genes are identified as significant genes by our method and are found to have significant effects in other three independent studies for validation. The most significant gene SOD2 has been identified in our previous osteoporosis study involving the same expression dataset. Several other genes such as TREML2, HTR1E and GLO1 are shown to be novel susceptible genes for osteoporosis, as confirmed from other studies.http://journal.frontiersin.org/Journal/10.3389/fcell.2014.00062/fullOsteoporosisGroup Lasso: Sparse regressionmultitask learningsignificant test |
| spellingShingle | Dongdong eLin Dongdong eLin Jigang eZhang Jigang eZhang Jingyao eLi Jingyao eLi Hao eHe Hao eHe hong-wen eDeng hong-wen eDeng Yu-Ping eWang Yu-Ping eWang Yu-Ping eWang Integrative analysis of multiple diverse omics datasets by sparse group multitask regression Frontiers in Cell and Developmental Biology Osteoporosis Group Lasso : Sparse regression multitask learning significant test |
| title | Integrative analysis of multiple diverse omics datasets by sparse group multitask regression |
| title_full | Integrative analysis of multiple diverse omics datasets by sparse group multitask regression |
| title_fullStr | Integrative analysis of multiple diverse omics datasets by sparse group multitask regression |
| title_full_unstemmed | Integrative analysis of multiple diverse omics datasets by sparse group multitask regression |
| title_short | Integrative analysis of multiple diverse omics datasets by sparse group multitask regression |
| title_sort | integrative analysis of multiple diverse omics datasets by sparse group multitask regression |
| topic | Osteoporosis Group Lasso : Sparse regression multitask learning significant test |
| url | http://journal.frontiersin.org/Journal/10.3389/fcell.2014.00062/full |
| work_keys_str_mv | AT dongdongelin integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT dongdongelin integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT jigangezhang integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT jigangezhang integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT jingyaoeli integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT jingyaoeli integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT haoehe integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT haoehe integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT hongwenedeng integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT hongwenedeng integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT yupingewang integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT yupingewang integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression AT yupingewang integrativeanalysisofmultiplediverseomicsdatasetsbysparsegroupmultitaskregression |