how to do Manhatton plot for significant DNA methylation sites associations

Question

[Bioinfonext]

Hi,

I do have a limma output file for significant association, could you please suggest how to plot Manhatton plot for these sites;

sites <- read.table("cpg.site.txt", header=T, sep=",")

head(sites)

                  logFC           AveExpr       t                  P.Value              adj.P.Val               B           UCSC_RefGene_Name   UCSC_RefGene_Group  UCSC_RefGene_Name
cg17944885  -0.45182214 2.558778887 -15.5883061 1.29449893946335e-42    9.91956414325614e-37    84.49688759 0   0   chr19
cg03546163  0.255316075 -0.802081399    12.08720314 1.21449935724625e-28    4.65326927233398e-23    53.61163067 FKBP5;FKBP5;FKBP5;FKBP5 5'UTR;5'UTR;5'UTR;5'UTR chr6
cg05165263  -0.185316318    0.081109579 -10.79707879    7.43220006052377e-24    1.89839695185951e-18    43.01154759 IRF2    Body    chr4

Many thanks,

Answer 1

[dthorbur]

You can't really plot a manhattan plot with just 3 loci. They typically describe a distribution of values across an entire contig/chromosome/genome. Otherwise, a manhattan plot is just a scatter plot organised by genomic distance on the x and a given value on the y.

Here is a tutorial on how to do them in R with and without ggplot2. Halfway down the page are a few examples specifically of manhattan style plots.

EDIT: I realise there is the command head(sites), but it only printed 3 rows, when the default is 6 in R, hence my concern about the number of data points.

[Bioinfonext]

thanks, I just put 3 loci here, but I do have some more in the output list.