Introduction to Bioinformatics - Biol 591             

 

SCENARIO
Distinguishing different classes of acute leukemia
Finding patterns in microarray data

Finding patterns in variation
It may seem a simple step, but it isn't.

There is considerable variation in gene expression from patient to patient, which confounds the analysis. However, if one considers dozens of patients and thousands of genes, it might be possible to discern a pattern from the sea of numbers. Which are the genes that are truly informative and how might their information be combined?

           Patients
 Gene    Pt1  Pt2  Pt3  Pt4  Pt5 ...
 M22538  704 4107 1010  905 2204 
 M22638  552  131 -288  336  677 
 M22760  625 1594 1257 1122 2318 
 M22877  275  557  585  244  941 
 M22898  474  339  650  613  873 
 M22919  332 1754  588  161  543 
 M22960  330  811  212  584 1018 
 M22976  369  379  578  487  436 
 M22995  424  609  214  245  559 
 ...
It is for these kinds of questions that statistics was invented, and it is evident that we must grasp something about statistical measures to understand how microarray data may be used in diagnosis.

How can we extract useful information from microarrays?

back      back to main page