Date

2013

Author

Akal, Tülay

Metadata

Show full item record

Item Usage Stats

151
views

43
downloads

Cite This

The microarray technology is one of the recent and advance tools in biological sciences. This optical technology aims to measure the amount of changes in transcripted message for each gene by RNA via quantifying the colour intensity on the arrays. But due to the different exper- imental conditions, these measurements can include both systematic and random erroneous intensities. In this study, we deal with one of these systematic sources of errors, called background sig- nals, for one-channel microarrays. Hereby, we initially describe the most well-know methods such as MAS 5.0, MBEI, RMA, and BGX approaches for estimating the gene expression levels, i.e., gene expression indices. Then, we present a novel gene expression index, called multi-RGX (Multiple Probe-Robust Gene Expression Index), which can be seen as a general- ization of the FGX model and closely related to the BGX method developed for this type of ar- rays. In multi-RGX, the FGX model is extended by both covering nonnormal log-expressions, in particular, long-tailed symmetric (LTS) densities, and taking not only the probe mean in- tensities, rather using all gene expressions in each probe for every gene. In inference of such model, we apply the modified maximum likelihood method to deal with the unexplicit so- lutions of the likelihood equations under LTS. Moreover, we derive the covariance-variance matrix of model parameters from the observed Fisher Information matrix. Finally in order to find the gain in information from the estimation, we evaluate the performance of our novel index in different datasets.

Subject Keywords

Gene expression., Information theory in biology., Oligonucleotides., Protein microarrays., Biometry .

URI

http://etd.lib.metu.edu.tr/upload/12615923/index.pdf
https://hdl.handle.net/11511/22630

Collections

Graduate School of Natural and Applied Sciences, Thesis