ratioplot {qpcR} | R Documentation |
A plot method in which data obtained from ratiocalc
is displayed as a barplot or hanging barplot. Error bars can be added to the bars
that had been calculated by any of the four different error values of propagate
, i.e.
propagated/Monte Carlo propagated/evaluated/Monte Carlo evaluated errors. Either significance codes or values from the t-tests can be added to
the bars.
ratioplot(data, errbar = c("prop", "propSim", "eval", "evalSim", "none"), sem = FALSE, order = NA, y.fac = 1.5, normcol = NA, errwid = NULL, type = c("bar", "hbar"), plot.t = c("none", "stars", "values"), offset = 1, ...)
data |
data of class 'ratiocalc'. |
errbar |
type of error that is used in displaying the error bars. See 'Details'. |
sem |
logical. Should the standard error of the mean (s.d./sqrt(n)) be displayed? |
order |
a numeric vector defining subsets or reordering of the plotted bars. |
y.fac |
multiplication factor for the y-axis extension, can be tweaked for a more appealing output. |
normcol |
column number that all other data should be normalized against. See 'Details'. |
errwid |
The width of the error bars. If NULL (default), calculated from the width of the bars. |
type |
the plot type. See 'Details'. |
plot.t |
how to display the results from the t-tests. Either significance codes (i.e. < 0.01 => "**") or the p-values. |
offset |
distance of plot.t from the end of the bars. For tweaking. |
... |
other parameters to be passed to barplot or arrows . |
Ratios are displayed from all permutations/combinations of sample/replicate sample PCR runs, as described in ratiocalc
. The data can also be normalized
against one of the runs, in which the 'control' data is normalized to 1 and all other data and their errors accordingly.
The following error types can be displayed:
prop
the propagated error (standard deviation);
propSim
the propagated error averaged from Monte Carlo simulation;
eval
the (unpropagated) error (standard deviation) resulting from averaging the expression evaluation;
evalSim
the (unpropagated) error (standard deviation) from the evaluated expressions of the Monte Carlo simulation.
If type
= "bar", a normal error bar plot is displayed. If type
= "hbar", ratios < 1 are displayed as hanging bars with (-1/value).
The latter is more appealing and visually discriminates the up-/down-regulated samples.
A barplot with error bars displaying the ratios obtained from ratiocalc
.
Andrej-Nikolai Spiess
## quick (but not dirty!) analysis of qPCR data! ## normal bar plot with standard error of mean ## and t-test on crossing points DAT <- modlist(reps, 2:9, fct = l5()) GROUP <- c(1, 1, 2, 2, 3, 3, 4, 4) res <- ratiocalc(DAT, group = GROUP, ttest = "cp") ratioplot(res, cex.names = 0.3, sem = TRUE, plot.t = "values") ## hanging bar plot with standard deviation ## and significance codes ratioplot(res, cex.names = 0.3, type = "hbar", plot.t = "stars") ## same using the 'unpropagated' error and ## different y-scale ## (from averaged expression evaluations) ratioplot(res, cex.names = 0.3, type = "hbar", errbar = "eval", ylim = c(-20, 59))