how to visualize the distribution of 20,000 genes in gene expression data

Question

[mms140130]

Hi,

I need to visualize the distribution of 20,000 genes for 1000 patients in gene expression data to see how far it is from Normal distribution assumption , to perform an eQTL analysis to see if the transformation of log is meaningful ( no bimodal plots)

Can anyone suggest how to visualize the distribution of 20,000 genes

[mforde84]

Why not just use plot() for each gene?

[mms140130]

is this an efficient way to see all plots at once , so I have 20,000 plot ??!

[mforde84]

I don't know if you're better off heatmapping a quarter million data points though, especially if you're not scaling the data for each gene. Maybe visual examination is out of the question. Instead iterate through each gene, and do a shapiro test for each to assess normality:

shapiro.test(y)$p.value

Combine these pvalues into a list for all of your genes, and sort by pvalue. Anything significant is a normal distribution.