- The columns Order and RColorCode in the configuration file are now optional.
Only columns SampleName and GroupName are mandatory.
This should simplify the minimum requirements needed to create a functional config.
- Without Order, the order defaults to 1,2,...,N where 1 is the first row of the config and N is the last row.
- Without RColorCode, the default ggplot2 color palette is used.
- New subgroup feature: instead of displaying the average PSI of subgroups, the individual point estimates are plotted.
- Updates to vignette
- Bug fixes
- Support for sample subgroups in all plot functions, through the config files. PSIs of samples in a subgroup are averaged and displayed as a single data point. Confidence intervals are computed by fitting a joint beta distribution to a population of points sampled from beta distributions coming from the individual samples, and then applying the same method as for single samples.
- Support for plotting multiple events in the same PSI plot, with events identified with different colours, and samples identified by the color background.
- Support for sample columns containing suffixes (e.g. "Sample1-cRPKM", while the config says only "Sample1").
- Support for gene expression tables containing both cRPKM and raw read counts (similar to the PSI and QC columns for events).
- Big changes in the preprocess_sample_colors function and its output, to allow for all this.
- Use of dplyr (and tidyverse in general) instead of plyr.
- README contents moved to vignette
- Fixed bug that occurs when dealing with columns that are all NA
- Added CITATION
- Converted some internally-used functions to be non-exported
plot_multi()
now supportspheatmap
- The
test_data
folder is nowinst/extdata
Some bug fixes for plot_expr()
This release contains new features and some internal updates. In addition, this release now requires R version 3.1 or higher and ggplot2 version 2.0 or higher.
- Replace
show_guides
withshow.legend
to match API of ggplot2 version 2.*. - Update README
plot_multi
has been further developed:- Use
heatmap.2
from gplots package to produce heatmap. To minimize dependencies, this package is optional and needs to be manually installed (install.packages("gplots")
). In the absence of gplots, ggplot2 will be used as before. - Events and samples can be clustered using options
cluster_rows
andcluster_cols
, respectively. - Add
fill
option for providing custom colours - Simply format of row names
- See
?plot_multi
for full set of options
- Use
- Retire warning for deprecated
lines
option - The option
plot
inplot_event
andplot_expr
is deprecated (will issue warning if used). - Other documentation updates
This release contains minor updates.
plot_event
: defaultylim
changed to (0,100) from (1,100).
Version 2.0.0 now uses ggplot2 to generate plots in plot_event
and
plot_expr
. In addition to outputting a plot, a ggplot2 object is returned.
- Printing of the plot can be suppressed with
plot = FALSE
. - The option
lines
inplot_event
andplot_expr
is deprecated (will issue warning if used). - Imports plyr for calculating group means.
plot_multi
: a new experimental function that will generate a heatmap of PSI values for two or more events:
plot_multi(psi)
- Change y-axis labels to be horizontal (e.g.
las = 1
)
- Changed minimum of y-axis to start at 0 for
plot_expr()
(#2) - Added check if config and input data does has 0 matching samples (#3)
- Other minor updates to documentation
plot_expr
: psiplot can now plot cRPKMs. For example, using the new sample datasetcrpkm
:
plot_expr(crpkm[1,], config = config)
- Renamed sample datasets to:
psi
andconfig
(#1) - Deprecated
xlim
argument inplot_event
- it shouldn't really be used
- This is the first release of psiplot in the form of an R package.