統計学輪講(第39回)

日時      2010年02月02日(火)    15時50分~16時40分
場所      経済学部新棟3階第3教室
講演者    Md. Manir Hossain MOLLAH (農)
演題      Beta-divergence approach as a tool of robustification and
diagnosis of EBarrays

概要

In a gene expression array experiment, the expression levels of thousands of
genes are monitored simultaneously. To reduce the dimensionality, Kendziorski
et al. (2003) proposed an empirical Bayes approach of hierarchical mixture
model (EBarrays) that accounts for differences among genes in their average
expression levels, differential expression for a given gene among cell types,
and measurement fluctuations. In this report, we robustify the EBarrays model
by adopting the beta divergence, the generalized measure of KL distance (Basu
et al., 1998; Minami and Eguchi., 2002). By simulation, we show that our
procedure manages the high true positive rate with  smaller false negative rate
compared with the other methods (Bridge (Gottardo et al., 2006) and GaGa
(Rossell, 2009)) when the data includes outliers. We also show that the weight
plot can be used  as a tool for model diagnosis. Application to the expression
data from the case-control study of head and neck cancer (Kuriakose et al.,
2004) confirmed the efficacy of our procedure.

 References

1. Basu, A., Harris, I.R., Hjort, N. L., & Jones, M. C. (1998). Robust and
efficient estimation by minimizing a density power divergence. Biometrika,
85: 549-559.

2. Gottardo, R., A. E. Raftery, K. Y. Yeung, and R. Bumgarner (2006).
Bayesian robust inference for differential gene expression in cDNA
microarrays with multiple samples. Biometrics 62: 10-18.

3. Kendziorski, C., Newton, M., Lan, H., and Gould, M. N. (2003). On
parametric empirical Bayes  methods for comparing multiple groups using
replicated gene expression profiles. Statistics in Medicine 22: 3899-3914.

4. Kuriakose MA Chen WT, He ZM, et al. (2004): Selection and
validation of differentially
expressed genes in head and neck cancer.  Cell Mol Life Sci. 61: 1372-1383.

5. Minami, M. and Eguchi, S. (2002): Robust blind source separation by
beta-divergence. Neural Computation, 14: 1859-1886.

6. Rossell, D. (2009). GaGa: A parsimonious and flexible model for
differential expression analysis. The Annals of applied Statistics, 3,
1035-1051.