Data Mining
RBM will help you mine the biomarker data generated using our MAP testing service. The first level of analysis identifies those analytes who demonstrate significantly different concentrations in experimental and control groups. Typically these are identified using means and standard t tests or analysis of variation (ANOVA). The significant analyte(s) are reported with their corresponding p values. Where necessary, these results are plotted using standard graphical plotting such as Excel. An example is shown below.
The second level of analysis utilizes a principal component analysis from BioWisdom (formerly OmniViz) Software of Maynard, MA. Using the significant analytes determined in the level one analysis, this software plots each sample in relation to the others in the study. This view is called the Galaxy™, and is very useful in visualizing the relationship of sample to one another. The steps in this analysis are typically: 1) normalization by computing the standard deviation around the mean for each significant analyte; 2) correlation of the movement of each analyte’s concentration to the other analytes; 3) pixilation of the correlation data creating a unique “picture” of each sample and finally, 4) the creation of the Galaxy view. An example of a Galaxy view is shown below.
This Galaxy view computed using the same six analytes shown above.