I want to analyze microarray data generated from beadstudio software using illumina. I want to know how do we decide whether our data will follow some pattern as in normal distribution or bayesian distribution or will not follow any trend.
Based on this, which statistical test one should use for normalization and differential gene expression?
My area of expertise is not genomics. Perhaps someone in the UCL genomics centre could advise.
Hi sabeena
R is a excellent and Flexible software for statistical analysis
Well,
certainly R is one of the most powerful programs to deal with this type of problem. But before to use the software, you'll know what statistical methods to use. In cDNA microarray expression data it is common to use T tests or SAM methods to detect differential expression. Maybe these may help you.
Just a quick note: 'R' is a package of statistical programs while 'r' is the usual symbol for a Pearson product-moment correlation coefficient. There are several tests of whether the distribution is 'normal': e.g., Kolmogorov-Smirnov, Anderson-Darling, and Shapiro-Wilk. Be careful: sample size and statistical power are concerns. Often the examination may be of thousands of genes and their expression, but the genetic initial data are drawn from a small number of individuals. In addition mto limitian stastistical power, this samping places profound limitations on generalizability (AKA external validity).
Sorry for typos:: the final sentence should read 'In addition to limiting statistical power. . .'.
Thank you all for sharing information. will try these test on my data set.
- Badges
- Science topic
- Similar topics
- Bioinformatics
- Bioinformatics and Computational Biology