A robust differential identification method that considers an ensemble of finite mixture models combined with a local false discovery rate (fdr) for analyzing ChIP-seq data comparing two samples. This package can also be used to identify differential in other high throughput data such as microarray, methylation etc. After normalization, an Exponential-Normal(k) or a Uniform-Normal(k) mixture is fitted to the data. The (k)-normal component can represent either differential regions or non-differential regions depending on their location and spread. The exponential or uniform represent differentially sites. local (fdr) are computed from the fitted model. Unique features of the package:
Using ensemble of mixture models allowing data to be accurately & efficiently represented. Then two-phase selection ensure the selection of the best overall model.
This method can be used as a general program to fit a mixture of uniform-normal or uniform-k-normal or exponential-k-normal
| Package: | DIME |
| Type: | Package |
| Version: | 1.0 |
| Date: | 2010-11-19 |
| License: | GPL-2 |
| LazyLoad: | yes |
Khalili, A., Huang, T., and Lin, S. (2009). A robust unified approach to analyzing methylation and gene expression data. Computational Statistics and Data Analysis, 53(5), 1701-1710.
Dean, N. and Raftery, A. E. (2005). Normal uniform mixture differential gene expression detection for cDNA microarrays. BMC Bioinformatics, 6, 173.
Taslim, C., Wu, J., Yan, P., Singer, G., Parvin, J., Huang, T., Lin, S., and Huang, K. (2009). Comparative study on chip-seq data: normalization and binding pattern characterization. Bioinformatics, 25(18), 2334-2340.