The Centre for Biostatistics, together with the Department of
Mathematics and Statistics, are pleased to announce the first of the
autumn term's seminars. Further details of the programme are available
at
http://www.ul.ie/biostatistics
Biostatistics Seminar
Friday, September 28th
2:00pm, Room A2002
Locating quantitative trait loci with a modified version of the Bayesian
Information Criterion
Malgorzata Bogdan
Wroclaw University of Technology, Poland
Purdue University, USA
We consider the problem of locating genes influencing quantitative
traits (Quantitative Trait Loci, QTLs). This task is usually
accomplished by using a multiple regression model and choosing the
important predictors with some model selection criteria. The Bayesian
Information Criterion, BIC, (Schwarz, 1978) is one of the most popular
model selection criteria. Due to the relatively large penalty for
including new components, BIC is often considered to be overly
conservative. However, as demonstrated by Broman and Speed (2002), in
the context of QTL mapping this criterion has a tendency to include many
spurious variables. This phenomenon was explained in Bogdan et al.
(2004), where a suitable modification of BIC, mBIC, was also proposed.
In this talk we will discuss the reasoning behind mBIC and demonstrate
its extension to rank-based statistics, designed to work with
`heavy-tailed' traits (Zak et al. (2007)). If time allows, we will also
present recent results on the adaptation of the method to genome scans
based on very dense sets of markers or interval mapping (Bogdan et al,
(2007)). The results presented on the behavior of model selection rules
are general and are relevant to any case of data mining.
References
Bogdan, M., Ghosh, J. and Doerge, R. W. Modifying the Schwarz bayesian
information criterion to locate multiple interacting quantitive trait
loci, Genetics 167 , 989-999, 2004.
Bogdan, M, Frommlet, F., Biecek, P., Cheng, R., Ghosh J. K.,
Doerge R. W. Extending the Modified Bayesian Information Criterion
(mBIC) to dense markers and multiple interval mapping, Biometrics,
accepted with revisions.
Broman, K. W. and Speed, T. P. A model selection approach for the
identification of quantitative trait loci in experimental crosses. J. R.
Stat. Soc. B, 64, 641-656, 2002.
Schwarz, G. Estimating the dimension of a model. Ann. Stat. 6,
461-464, 1978.
Zak, M., Baierl, A., Bogdan M., Futschik A. Locating multiple
interacting quantitative trait loci using rank-based model selection'',
Genetics, 176: 1845-1854, 2007 .
|