plot_event: Plot PSI values of a given alternative splicing event
In kcha/psiplot: Generate plots of PSI values generated by vast-tools

plot_event

R Documentation

Plot PSI values of a given alternative splicing event

Description

Generate a plot of PSI values for a given exon. The PSI values and corresponding quality scores are typically obtained from the vast-tools pipeline.

Usage

plot_event(x, config = NULL, subg = FALSE, subg.show = c("mean",
  "all", "beeswarm"), trim_colnames = NULL, qual = c("VLOW", "N",
  "LOW", "OK", "SOK"), errorbar = TRUE,
  groupmean = ifelse(is.null(config), FALSE, TRUE), col = NULL,
  title = NULL, xlab = "", ylab = "PSI", ylim = c(0, 100),
  cex.main = 14, cex.yaxis = 12, cex.xaxis = 12, pch = 20,
  cex.pch = 3, plot = NULL, gridlines = TRUE,
  show_group_legend = TRUE)

Arguments

`x`	A 1-row data frame containing PSI values to be plotted.
`config`	Optional configuration settings for `plot_event`. Can be a path to the `.config` file, or 4/5-column data frame of the `.config` file. Use the latter option if you are calling `plot_event` multiple times.
`subg`	Logical indicating whether samples should be subgrouped for plotting.
`subg.show`	Only applies when `subg == TRUE`. Default is `mean`, in which the average PSI is computed for each subgroup. If `all`, then individual point estimates with error bars are shown. If `beeswarm`, this is similar to `all`, but shown as a beeswarm plot and without error bars.
`trim_colnames`	String that must be searched for and trimmed at the end of every sample column in x. If no string must be trimmed, leave as `NULL`.
`qual`	String indicating the minimun vast-tools quality score for the PSI to be accepted. Defaults to `'VLOW'`. See the vast-tools documentation for details.
`errorbar`	Logical indicating whether error bars should be drawn.
`groupmean`	Logical indicating whether grouped means should be drawn. Requires `config`.
`col`	Vector of colors with length matching the number of samples. If specified, this will override the color settings specified in `config`.
`title`	Title of the plot. If `NULL` (default), the title generated by `make_title`.
`xlab`	The x-axis label.
`ylab`	The y-axis label.
`ylim`	Range of y-axis.
`cex.main`	Plot title size (pts).
`cex.yaxis`	Y-axis font size (pts).
`cex.xaxis`	X-axis font size (i.e. the sample names) (pts).
`pch`	Point symbol.
`cex.pch`	Size of datapoints.
`plot`	(deprecated) Prints the plot.
`gridlines`	Logical indicating whether grid lines should be drawn.
`show_group_legend`	Set to FALSE to avoid showing a legend with the sample groups and their colors.

Details

Plots can be customized via the config option. Either a data frame or the filepath to the config file can be used. Alternatively, plots can be customized using a limited set of graphical parameters as described above.

The order of samples (e.g. columns in x) as it appears on the resulting plot can be customized using a config file. If a config file is not used and re-ordering is desired, then it must be done manually before calling plot_event by altering the columns of x.

If subgroups are defined in config and subg=TRUE:

The mean PSI of the samples within each subgroup are drawn as a single data point.
If subg=TRUE and errorbar=TRUE, confidence intervals for the subgroup are derived from a joint beta distribution, fitted to a set of points derived from individual beta distributions corresponding to each sample (see get_beta_ci_subg). This is an experimental feature, computationally expensive, and CIs may not be shown for some subgroups,due to failure in fitting the joint distribution, especially for events with low coverage and PSI values near 0 or 1.
Subgroups will be ordered by the minimum Order value of their samples, and assigned to the first group to which they are matched.
(Experimental) Instead of computing the average PSI, the individual PSI values can be plotted for each subgroup by setting subg.show. The default is "mean", which plots the average PSI. The other two options are "all" or "beeswarm". For "all", the individual PSI values and error bars are plotted for each subgroup and sorted in increasing order. For "beeswarm", the optional ggbeeswarm package is used to create beeswarm scatterplots for each subgroup (error bars are not possible). As this is an optional feature, please manually install the package first before using it: install.packages("ggbeeswarm").

If no subgroups are defined in config, or subg=FALSE (default), a subgroup will be defined for each sample, preserving their name and order.

If groups and colors are defined in config, all the samples in a group will be colored with the first color assigned to the group. Else, all samples will be plotted in black.

If groups are defined in config and groupmean=TRUE, the mean PSI of the samples within each group are drawn as horizontal lines. The colors of the lines are determined by the color set to the first sample of each group by RColorCode in config. A corresponding legend key will also be drawn.

(See also preprocess_sample_colors and vignette('psiplot-usage') for details on how to use config files).

PSI values that have NA value are omitted and not plotted.

Error bars based on the confidence interval of the PSI estimation can be shown by setting errorbar=TRUE.

Value

ggplot2 object

Examples

## Not run: 
plot_event(psi[1,])
# Plot with custom configuration
config
plot_event(psi[1,], config = config, groupmean=TRUE)
plot_event(psi[1,], config = "/path/to/config")

# Plot with subgrouped samples
config
plot_event(psi[1,], config = config, subg = TRUE, errorbar = FALSE)

# Plot with subroups and error bars (slow, experimental feature)
\dontrun{
plot_event(psi[1,], config = config, subg = TRUE, errorbar = TRUE)
}

# Plot using custom configuration, changing point symbol, and y-axis
# scale
plot_event(psi[1,], config = config, pch = 9, ylim = c(20, 80))

## End(Not run)

kcha/psiplot documentation built on March 27, 2022, 4:20 a.m.