| Microarray array technology has fast become a powerful and popular tool for investigating high throughput gene expression in a variety of biological settings. The whole genome tiling microarray is a special type of microarray platform that allows uniform and unbiased interrogation of nearly the entire transcribed genome (transcriptome) with high resolution. In comparison to other array platforms, tiling arrays also present unique challenges in data analysis including larger data sizes per experiment, more variable probe behavior, and correlation among neighboring probes. Here, three different tiling array analyses are presented: (1) image quality control (2) prediction of transcribed regions and (3) quantification of known expressed regions. Before performing any specific microarray analysis, imaging quality control is assessed by a probability-based tool called MBR that both detects and removes image defects, resulting in improved downstream analysis results. Next, a correlation-based sliding window algorithm is used to predict regions of mRNA expression from eight tumor cell lines by leveraging all of the data simultaneously. Finally, a gene variant expression index, similar to that calculated in annotation-dependent expression arrays, is presented for quantifying transcript levels using high-density tiling microarrays. |
