[R] problem plotting points based on different values

[Dennis Murphy]

Hi:

A few things that are worth considering in generating a plot of this type:

(1) I suspect you would want larger point sizes with smaller p-values.
(2) To generate intervals of p-value categories, it's easier to
generate a factor in the data frame.
(3) It would probably make sense to have some sort of 'nice' color
gradient to associate with the p-value intervals.

To that end, I generated a small data frame and tried the following.
This assumes you can live with strict inequalities in the p-values.

dd <- data.frame(x = rnorm(30), y = rnorm(30), p = rexp(30, 30))
table(dd$p)   # see what you've got for p-value ranges
# Generate a set of p-value intervals using the cut() function:
dd$pval <- cut(dd$p, c(0, 0.0001, 0.001, 0.05, 1))
# Reverse the levels and generate some labels for the legend
dd$pv <- factor(dd$pval, levels = rev(levels(dd$pval)),
   labels = c('p > 0.05', '0.001 < p < 0.05', '0.0001 < p < 0.001',
              'p < 0.0001'))

# plot
# scale_colour_manual() is used to select a specific color scheme
ggplot(dd, aes(x = x, y = y)) + theme_bw() +
  geom_point(aes(size = pv, colour = pv)) +
  labs(size = 'p-value', colour = 'p-value') +
  scale_colour_manual(breaks = levels(dd$pv), values = topo.colors(10))

If you want the <= values in the legend in a nice 'mathematics-like'
inequality, that's not quite so easy...

Just for fun, here's another take on the plot above, pulling the
legend inside the plot region to increase the plotting area:

ggplot(dd, aes(x = x, y = y)) +
  geom_point(aes(size = pv, colour = pv)) +
  labs(size = 'p-value', colour = 'p-value') +
  scale_colour_manual(breaks = levels(dd$pv), values = topo.colors(10)) +
  opts(legend.position = c(0.85, 0.88),
       legend.background = theme_rect(fill = 'white',
                                      colour = 'white'))

If you don't like the white background on the legend, take out the
legend.background = statement in its entirety. Its advantage, though,
is that it 'paints over' several of the grid lines.

HTH,
Dennis


On Tue, Jul 12, 2011 at 4:34 PM, Hrishi lokhande <hrishi27n at gmail.com> wrote:
> Hello Friends,
>
> I am new to R and stuck with a problem.
>
> i have two columns drug_A and drug_B, i have plotted a scatter plot
> using the ggplot2 function.
>
> My problem is with the third column, it is the p-value column.
>
> I want to color and size points differently based on the p_value, the
> p_value range is between 0<0.0001< 0.001<0.05<1.
>
> I used a script using the ifelse loop, but it considers only two
> conditions at one time.
>
> Please tell me a way to fulfill all the conditions.
>
> here is the script:
> library(lattice)
>
>  qplot(Drug_A, Drug_B, data=dsmall, size = p_value, colour =
> ifelse(p_value > 0, ">0", "<0.0001"),
> + ifelse(p_value1 > 0.001, ">0.001", "<0.01")
> + ifelse(p_value1 > 0.001, ">0.001", "<0.05"
> ))
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>