Hotel Estoril Eden, Monte Estoril,
5-8 October 2005



NextText Box: Participants
Text Box: Programme

ew Softwares for Probe-Level Data Analysis: oligo and makePlatformDesign

Benilton Carvalho1, Rafael A. Irizarry1, Ben Bolstad2 and James W. MacDonald3
Department of Biostatistics, Johns Hopkins University, Baltimore, MD, USA
Department of Biostatistics, UC Berkeley, CA, USA
Affymetrix and cDNA Microarray Core, University of Michigan, Ann Arbor, MI, USA

Microarray is a well-known type of high through-put technology for genome-wide expression analyses. It has become a very popular tool in medical and biology researches. Microarrays are applied in a wide variety of DNA and mRNA analyses (for example, the elucidation of interactions between signaling pathways involved in development, the discovery of a new class of tumor, or the development of new assays to track drug metabolism) by measuring diverse high genomic endpoints including yeast mutant representations, the presence of SNPs, presence of deletions/insertions, and protein binding sites by chromatic immunoprecipitation (known as ChIP-chip). In each case, the genomic units of measurement are short oligonucleotides referred to as probes. Without appropriate understanding of the bias and variance of these measurements, biological inferences based upon probe analysis will be compromised.

Transforming raw data into genomic-level measurements requires a number of important steps. These data manipulations referred to as pre-processing, have enormous influence on the quality of the ultimate measurements and on the studies that rely upon them. Not only in this technology, much effort is spent on pre-processing. It has been demonstrated that the use of statistical methodologies can substantially improve accuracy and precision of expression measurements, relative to ad-hoc procedures introduced by designers and manufacturers of the technology. Although a large number of tools exist for the analysis of expression measurements, software for the analysis of probe-level data is quite limited. The further improvement of pre-processing procedures is an evolving research field and requires the availability of appropriate software.

Recently, expanding the features and flexibility offered by the Bioconductor affy package, two new packages were made available through Bioconductor (devel-branch): oligo and makePlatformDesign. Together, they aggregate the affy capabilities and support for different platforms (Affymetrix and NimbleGen), offering a flexible environment for probe-level data analysis, allowing contributed add-ons and improvements from users. The goal is to improve the software functionality and usefulness to the research community, therefore feedback is greatly appreciated.