Analysis parameters

Standardization is a hot topic in microarray data analysis. Microarrays generate so much data, which may be shared among many labs worldwide, that standardized experimental and analysis methods are becoming more important to the microarray community. When the issue of standardization is raised, conversation quickly turns to MIAME (Minimum Information About a Microarray Experiment) (see also Chapter 22) (5) and MAGE-ML (Microarray Gene Expression Markup Language) (6). MIAME is a guide to the types of information that scientists should record and report when describing microarray experiments. MAGE-ML is a file format for microarray data that contains MIAME descriptors in the file. The intended advantage of MAGE-ML over other file formats is that anyone can look at a MAGE-ML file and determine how the data was analyzed.

MIAME and MAGE-ML are successful in their intent, but they are a solution to only a very small part of a very large problem. Knowing how a microarray image was analyzed does not answer the much more important question: What is the best way to analyze a microarray image? What is the best segmentation method, the best background subtraction method, the best ratio measure, or the best normalization method? Do you even need to do all of these? The microarray community does not agree on an answer to any of these questions. This is a

Plate 1.

Classification of renal cell carcinoma samples using prediction analysis for microarrays (PAM) (21). The frequencies of correct tumor classification (y-axis) are plotted for each tumor sample (x-axis; green: ccRCC; grey: pRCC; orange: chRCC). The sample that is incorrectly classified in all repetitions is marked with a red triangle, all others with blue circles.

Plate 2.

nil mm

0 0

Post a comment