|
[Overview |
Imputation of missing values | Applications| Compression of Microarray Images]
Genomic data present many challenges for the statistician; there are missing values, the data are noisy and high-dimensional, and the biological context is important but not always easy to account for in a controlled fashion.
I am developing statistical methodology for missing value imputation, and the joint analysis of micro-RNA and mRNA expression (see below), and clustering and gene selection (see the The Clustering Research Page ). Recently, Sunduz Keles and I, in collaboration with our post-doc Roxana Alexandridis, started a new project, focused on incorporating biological information into gene expression analysis.
Imputation of missing values:
| In collaboration with Ming Ouyang, I am working on the development of missing value imputation tools for microarray data.
|
Recently, our Meta-Data based approach was accepted for publication in BMC Bioinformatics. We use the extensive public resources currently available to boost the power of a particular experiment. Thus, we downloaded all the SMD microarray data sets, and constructed a learning algorithm that imputes missing values in a query data set from the public data sets. We show that the accuracy of imputation is vastly improved, and we present a web-based tool for application.
| These days we are getting many requests for the code of our
Adaptive Microarray Data Imputation tool . This imputation scheme adapts to the characteristics of the data to select the most appropriate imputation scheme. The matlab code is available upon request.
|
|
Applications:
| I have an ongoing collaboration with The Hart Lab at Rutgers University. Currently, our research is focused on the joint analysis of miRNA and mRNA expression changes during neurogenesis. We are also working on a stem-cell project with Invitrogen.
|
Ming Ouyang and I are in the process of developing computational tools for assessing miRNA-mRNA target predictions using miRNa and mRNA expression and sequence data.
Compression of microarray images: Microarray images are but the basic building block of genomic analysis. Enormous quantities of image data are now collected in scientific labs throughout the world. For simple storage and fast transmission, image compression tools can assist. Several papers on the compression of microarray images can be found under Publications.
|
Clustering of high-dimensional biological data
Please visit the The Clustering Research Page
Microarray Data Imputation
Adaptive missing value imputation Slides from the 2005 NSF-RCN Retreat
Development, Evaluation, & Dissemination of Methods for the Analysis of
Gene Expression by Microarrays
Sept 19 - 23, 2005
DNA microarray data imputation and significance analysis of differential expression
Rebecka Jörnsten, Hui-Yu Wang, William J. Welsh and Ming Ouyang
Bioinformatics 2005 21(22):4155-4161
A Meta-data based method for DNA microarray imputation.
Rebecka Jörnsten, Hui-Yu Wang and Ming Ouyang
To appear in BMC Bioinformatics
In collaboration with the Hart lab
Screening anti-inflammatory compounds in injured spinal cord with
microarrays: A comparison of bioinformatics analysis approaches.
Jonathan Z. Pan, Rebecka Jörnsten and Ronald P. Hart
Physiol. Genomics (February 17, 2004).
Neuroscience meeting poster 2006
More posters and conference presentations can be found at The Hart lab home page
Compression of Microarray images
Comprestimation: Microarray images in abundance, Proc. of Conference on Information Science and Systems, Princeton, 2000
Microarray image compression: SLOCO and the effects of information loss.
Rebecka Jornsten, Bin Yu, Wei Wang and Kannan Ramachandran
EURASIP Signal Processing Journal, Special issue on genomic signal processing, (2002), 83/4, 859-869
Microarray image compression, slides from IMS 2002
On the bit-plane compression of Microarray images
Rebecka Jornsten, Yehuda Vardi and Cunhui Zhang
Statistical Data Analysis based on the L1-norm and Related methods. Birkhauser Statistics for Industry and technology, Y. Dodge editor.
Compressing Genomic and Proteomic Array Images for Statistical Analysis.
Rebecka Jornsten and Bin Yu
Invited book chapter Genomic Signal Processing and Statistics. (2004) E.R. Dougherty, I. Shmulevich, J. Chen, Z.J. Wang Editors.
|
Tutorials
slides from the NIAMS short course on microarray data analysis, Atlanta, March 6-7
Significance analysis, Imputation, Classification and Clustering.
slides from the NIDDK short course on microarray data analysis, San Francisco, March 30-31, 2006
Classification, Clustering and Validation
slides from the IAS seminar Institute for Advanced Sciences, 2004
Links
Visit
The Hart Lab at Rutgers University.
Visit
ebCTC, environmental bioinformatics and Computational Toxicology Center. This Consortium of UMDNJ, Rutgers University and Princeton University is funded by USEPA STAR Grant number GAD
R 832721-010.
|
