ggplot2, dplyr, and tidyr.size usage to linewidth where required by recent ggplot2.size -> linewidth migration for the remaining line layers that
still passed the deprecated size argument: the mean/median reference line added
by gghistogram(add = ...) and ggdensity(add = ...), and the connector segments
of ggdotchart(add = "segments"). These no longer emit ggplot2's "size aesthetic
for lines was deprecated" or "Ignoring empty aesthetic: size" warnings, and the
requested line width is now applied via linewidth.ggmaplot() no longer emits an "Ignoring empty aesthetic: size" warning on its
default call; the point layer sets size only when the user supplies a value.clean_lock_files() helper.ggrepel dependency to >= 0.9.2 to keep Ubuntu oldrel
(R 4.4.x) CI dependency resolution working.stat_compare_means() and geom_pwc().geom_pwc() so users can see which grouped
subsets were skipped and why (e.g., missing ref.group or insufficient levels).geom_pwc() with an explicit list of comparisons no longer drops a whole facet/panel
when a single requested pair cannot be tested (e.g. a group that is entirely NA or has
fewer than two observations). The untestable comparison is skipped (with a message) and
the remaining valid comparisons are still drawn (#542).geom_pwc() now detects an absent ref.group in a grouped subset via the
rstatix_missing_ref_group condition class raised by recent rstatix
(walking the parent chain with rlang::cnd_inherits()), in addition to
matching the error message. This keeps the "skip ref-less subsets" behaviour
working after rstatix made that error message clearer (rstatix #153); older
rstatix versions are still handled via the message fallback.format_p_value(), get_p_format_style(), and list_p_format_styles().default, apa, nejm, lancet, ama, graphpad, scientific).p.format.stylep.digitsp.leading.zerop.min.thresholdp.decimal.markp.format.signif support and related label handling paths.stat_cor() gains two label-formatting arguments:
r.leading.zero — set to FALSE to drop the leading zero of the
correlation coefficient (e.g. .73 instead of 0.73), completing
APA-style reporting together with p.leading.zero (#540). The dropped
leading zero is preserved through plotmath rendering (the value is quoted
so it is not silently re-normalized back to 0.73 in the default
expression output).p.coef.name — symbol for the p-value label; use "P" for an uppercase
p-value (#541).Passing ggtheme = NULL to the plotting functions (e.g. ggscatter(),
ggboxplot(), ggline(), gghistogram(), ...) now skips applying a ggpubr
theme, so the plot keeps ggplot2's default theme or the theme set globally with
theme_set(). Previously an explicit ggtheme = NULL was treated like an unset
argument and theme_pubr() was still applied. Calls that omit ggtheme or pass
a specific theme are unchanged (#561).
ggarrange() gains a spacing argument to increase the gap between the
arranged plots (a long-requested option). It adds a uniform margin, in
text-line units, around each plot. Default is 0 (no extra space; each plot
keeps its own margins, so existing arrangements are unchanged) (#151).
ggbarplot() gains a numeric.x.axis argument (already available in
ggline() and ggerrorplot()). With numeric.x.axis = TRUE the x variable
is kept numeric instead of being coerced to a discrete factor, so bars are
drawn at their numeric x positions (e.g. a time axis). Default is FALSE
(unchanged behavior) (#463).
geom_bracket() and stat_pvalue_manual() gain a tip.length.ref argument
controlling what tip.length is a fraction of. With the default
tip.length.ref = "data" the tips are a fraction of the data range (unchanged
behavior). With tip.length.ref = "axis" the tips are a fraction of the
y-axis range (ylim/scale_y_*), which renders at the same physical fraction
across plots and therefore gives visually constant tip lengths regardless of
the data range - useful to keep tips consistent across facets or across
separate plots with different scales (#362). stat_compare_means(comparisons =)
draws its brackets via ggsignif::geom_signif() and does not support
tip.length.ref; passing it there now emits an informative message pointing to
stat_pvalue_manual(..., tip.length.ref = "axis") instead of being silently
ignored (#362).
stat_cor() and stat_regline_equation() gain a label.y.step argument giving
the vertical spacing (in text-line units) between the labels of successive
groups. Default is 1.4 (unchanged behavior). Setting label.y.step = 0 stops
the per-group vertical shift so labels align across facet panels when a factor
is mapped to an aesthetic that also defines the facets (the long-standing
"labels climbing stairs across facets" behavior; #248). This control was
inspired by the vstep argument in the ggpmisc package (by Pedro J. Aphalo),
from which ggpubr's correlation/regression label-positioning logic was
originally adapted; this is acknowledged via @seealso in ?stat_cor.
ggviolin() gains a drop argument (passed to ggplot2::geom_violin()),
default TRUE (unchanged behavior). Previously the argument was silently
dropped, so grouped violins where a sub-sample has too few points to draw a
density (which geom_violin() removes, including from the dodge position)
could not be kept aligned with added boxplots / dot plots. Setting
drop = FALSE together with position = position_dodge(0.8, preserve = "single")
now reserves the empty dodge lane so all geoms stay aligned (#381).
geom_bracket() and stat_pvalue_manual() now place brackets correctly on a
transformed y axis (e.g. scale_y_log10()). y.position is given in data units
but was previously used directly in the scale's transformed space, so brackets
landed far off (e.g. at 10^30) and squashed the plot. The bracket y.position
is now run through the scale's transformation; on an untransformed (identity)
scale this is a no-op, so existing plots are unchanged (#342).
ggboxplot() now accepts a position argument (e.g.
position = position_dodge(0.9)), like ggviolin()/ggdotplot() already do.
Previously passing position errored (formal argument "position" matched by multiple actual arguments) because the dodge was hardcoded; the default is
unchanged (position_dodge(0.8)), so grouped box plots look the same unless
you set it (#615).
The core plotting functions (ggscatter(), ggboxplot(), ggviolin(),
ggline(), ggbarplot(), gghistogram(), ggdensity(), ggdotplot(),
ggstripchart(), ggerrorplot(), ggecdf(), ggdotchart(), ggpaired(),
ggqqplot()) now accept the British spelling colour as an alias for
color. Previously colour was silently ignored (#317).
stat_regline_equation() now displays the correct equation for orthogonal
polynomial fits such as formula = y ~ poly(x, 2). Previously the orthogonal
basis coefficients were printed as if they were raw polynomial coefficients,
giving a wrong equation that did not match the fitted curve. For a simple
poly(x, k) term with an intercept, the equation is now computed from an
equivalent raw = TRUE fit (identical curve, R², AIC, BIC). Linear,
raw = TRUE, and I(x^2) formulas are unaffected; no-intercept models and
transformed poly arguments (e.g. poly(log(x), 2)) keep their previous
behavior (#653).
ggscatterhist() now aligns the marginal plots with the main scatter plot's
axes. The margins were built from the raw data, so anything that changed the
scatter limits (ellipse = TRUE, position jitter, explicit xlim/ylim)
left the marginal histograms/densities misaligned with the scatter. Each
margin now inherits the scatter's axis range (#220, #420).
geom_pwc() now places comparison brackets over the correct groups when an
entire x-axis group has only NA values. Such a group is dropped before the
statistical test runs; previously the surviving groups were renumbered from 1,
which shifted the remaining brackets left. The bracket positions are now
anchored to the discrete x scale so they stay aligned with the plotted groups
(#575).
Fixed ColorBrewer sequential palette parsing in .get_brewer_pal() so
"YlOrBr" is recognized correctly.
Updated ggexport() to respect verbose = FALSE by suppressing filename
print() output in multi-file raster/vector exports.
Hardened format_p_value() by validating non-NULL p.min.threshold as a
single positive finite number.
Updated create_p_label() to preserve NA values in p.format (returning
NA_character_ instead of stringifying to "p = NA").
ggballoonplot() now honors user-supplied xlab/ylab instead of always
blanking the axis titles; axis titles are still blank by default (#639).
ggsummarytable() gains an angle argument to rotate the summary-table text;
default (angle = 0) is unchanged (#595).
ggmaplot() gains a line.color argument to set the threshold line color;
default ("black") is unchanged (#322).
ggpar(legend.title = ) now also titles size and alpha legends, not just
colour/fill/linetype/shape (#412).
facet()/panel.labs: a NAMED panel.labs vector is now matched to the data
levels by name, fixing mislabeled panels when the order differed; unnamed
labels still map positionally (#643).
stat_compare_means() no longer fails with a "*.npc coord ..." error when a
grouped x value is 0 (or negative); .group_coord() now guards against a
non-positive group index and falls back to the first label coordinate (#594).
ggpar()/plot functions now apply xticks.by and yticks.by together; an
internal else if previously dropped xticks.by whenever yticks.by was
also supplied (#333).
stat_compare_means() now forwards the family argument to the comparison
bracket labels; it was previously ignored whenever comparisons was set
(#592, #624).
ggline() now supports two grouping variables at once (e.g. color and
linetype, or an explicit group plus color); these are combined into a
single interaction group so the right points are connected. Previously this
errored with "the condition has length > 1" (#616, #375).
A NAMED palette vector no longer emits a spurious "No shared levels found
between names(values) ..." warning; the manual color/fill scale is now only
applied to an aesthetic actually mapped in the plot. Unnamed palettes are
unchanged, and a named palette whose names genuinely don't match still warns
(#642).
ggviolin() no longer crashes with scales = "free" (or "free_x"/"free_y").
The facet argument scales was being partial-matched to the violin scale
parameter; it is now read with exact matching, so scales controls faceting
and the violin scale keeps its default (#398).
xlim/ylim are now honored with orientation = "horizontal" (and
rotate = TRUE). The limits are passed to coord_flip() instead of a separate
coord_cartesian() that was silently replaced, which also removes the
"Coordinate system already present" warning (#646).
ggpar() (and the plot functions) now honor legend.direction for all legend
positions; it was previously ignored for legend = "top"/"bottom" (#652).
compare_means() no longer errors ("Can't extract columns that don't exist")
when a formula variable name contains a space (e.g. len ~ `spa ced` ); the
backticks that R adds to non-syntactic term names are now stripped before
matching data columns (#385).
ggarrange() gains a byrow argument (default TRUE) to fill the plot grid
by column (byrow = FALSE) instead of by row; forwarded to
cowplot::plot_grid() (#225).
stat_regline_equation() gains coef.digits and rr.digits arguments to
control the number of significant digits shown for the regression-equation
coefficients and R2; defaults (2) reproduce the previous output (#312).
table_cell_font() and table_cell_bg() can now style an individual header
cell (row = 1), not only body cells; they previously matched only the
core-* grobs and silently did nothing for the header (#535).
theme_pubr() now draws the axis tick marks in black with linewidth 0.5,
matching the axis lines; previously the ticks inherited a lighter grey/thinner
style, visibly inconsistent when zoomed (#668).
theme_pubr() now sets strip.clip = "off" so the facet strip background
border renders at its full linewidth. With the ggplot2 (>= 3.5.0) default
(strip.clip = "on") the border was clipped to the strip area, cutting the
outer half of the stroke so it looked thinner and misaligned with the panel
when zoomed (follow-up to #668). Note: a facet label wider than its panel now
overflows the strip instead of being truncated; restore clipping with
+ theme(strip.clip = "on") if needed.
xticks.by/yticks.by now anchor the axis breaks to round multiples of the
step (e.g. 0, 100, 200) instead of the slightly-negative expanded axis minimum,
which produced odd labels such as -20, 80, 180 on bar plots with
ylim = c(0, 400) (#313).
ggmaplot() now draws the non-significant ("NS") points behind the significant
ones, so the up/down-regulated hits are no longer hidden under the grey NS
cloud; the legend order (Up, Down, NS) is unchanged (#365).
compare_means() now adjusts p-values within each group.by level (and
response) rather than pooling all groups together, so a grouped adjustment
matches filtering to one group and adjusting there. This changes p.adj values
for calls that use group.by; ungrouped calls are unchanged (#200).
stat_compare_means(comparisons = ) now emits a one-time message (and the docs
note) clarifying that the displayed pairwise p-values are not adjusted for
multiple comparisons, pointing to geom_pwc() / stat_pvalue_manual() +
compare_means(p.adjust.method = ) for corrected p-values (#293).
ggboxplot() now forwards coef to geom_boxplot(), so coef = 0 can be used
to omit the whiskers; it was previously dropped by geom_exec() (#517).
Import the %||% operator from rlang; it is used in geom_bracket() and
geom_pwc() but base R only provides it since R 4.4, so it could be unresolved
on the R (>= 4.1) versions the package supports (#665).
tab_add_title() / tab_add_footnote(): the just argument now actually
positions the text — "center"/"right" anchor the title/footnote across the
table width instead of around a fixed point — and the text is no longer clipped
when it is wider than the table (#302).
annotate_figure(): the figure label (fig.lab) now uses a length-independent
horizontal justification, so labels of different lengths keep the same anchor
and captions align across figures; previously a longer label was shifted away
from the corner (#185).
stat_compare_means() fatal warning/error in sparse grouped subsets with modern tidyr).stat_compare_means(label = "p.format") rounding and formatting control).reframe(), slice_head(), slice_tail(), across(), and where() functions.linewidth parameter to ggboxplot(), gghistogram(), ggviolin(), and ggdensity() for ggplot2 3.4.0+ compatibility. The size parameter is deprecated for line width in these functions (#644, #645, #654, #656, @erdeyl).adjust parameter to ggviolin() to control bandwidth adjustment for kernel density estimation (#552, @erdeyl).bw and adjust parameters to ggdensity() for bandwidth control (#490, @erdeyl).stat_cor(), stat_compare_means(), and stat_regline_equation() now use after_stat() syntax instead of deprecated ..var.. notation in default_aes (#645, @erdeyl).ggballoonplot() example updated to use guides(size = "none") instead of deprecated guides(size = FALSE) (@erdeyl).tidyr::gather() with tidyr::pivot_longer() in ggballoonplot() and compare_means() internals (#536, @erdeyl).dplyr::do() with dplyr::reframe() in compare_means(), desc_statby(), and internal helpers. Replaced dplyr::mutate_if() with dplyr::mutate(across(where())) in ggsummarytable() (@erdeyl).border() deprecation warning by using linewidth instead of size in element_rect() (#644, #654, #656, @erdeyl).size deprecation warnings in ggscatter() (rug and star plots), ggpaired() (connecting lines), ggecdf() (ECDF line), ggdensity() (density lines), geom_bracket(), and geom_pwc() (#645, @erdeyl).stat_cor() parsing error when options(OutDec = ",") is set (European decimal separator) by using decimal.mark = "." in formatC() calls (#512, @erdeyl).compare_means() error "object 'group2' not found" when using ref.group with method = "anova" or method = "kruskal.test" (#572, @erdeyl).exact = FALSE workaround from version 0.6.2 that forced non-default behavior on wilcox.test(). Tests now use flexible assertions to ensure compatibility across R versions (#649, #647)..parse_font() not recognizing decimal font sizes (e.g., lab.font = c(2.4, "italic", "black")), which caused label colors to render incorrectly (#659).compare_means() function now sets exact = FALSE for wilcox.test() and pairwise.wilcox.test() to maintain backward compatibility and consistent p-values across R versions (#647).outliers parameter to ggboxplot() to control the display of outlier points. Set outliers = FALSE to remove the black dots representing outliers from box plots (#614, @hswl1314)...p.signif.., ..eq.label..) to modern after_stat() calls with proper namespace qualification.ggline() parameter handling for ggplot2 3.4.0+ compatibility:
linewidth parameter for line widthsize parameter for lines with helpful warning messagesize parameterafter_stat() calls that were causing failures in reverse dependency packages (bSi, PopComm). The convert_label_dotdot_notation_to_after_stat() function now properly handles namespace qualification while maintaining backward compatibility (#638).ggplot2::after_stat() is accessible during plot building, resolving "could not find function after_stat" errors in downstream packages.stat_regline_equation() to display in standard mathematical convention "y = mx + b" instead of "y = b + mx" (#559, @tshates, @mwaak).gghistogram() tests to handle changes in binning standardization introduced in ggplot2 4.0.0 (#635, @teunbrand).stat_pvalue_manual() failing when fill or other aesthetics are provided in the parent ggplot layer. The function now sets inherit.aes = FALSE by default to prevent conflicts between parent plot aesthetics and the p-value annotation data (#621, @fncokg).ggplot2::is.ggplot() with ggplot2::is_ggplot() in ggpar().data$column syntax to quoted column names in geom_pwc() for tidyselect 1.2.0+ compatibilityall_of() wrapper in unnest() utility function for tidyselect compatibilitysize by linewidth in ggplot2 element_line() and element_rect() functions.stat_regline_equation() by automatically converting deprecated dot-dot notation (..eq.label.., ..adj.rr.label.., ..p.signif.., etc.) to after_stat() syntax for ggplot2 3.4.0+ compatibility (#623, @hinkyisme).add_summary() and ggerrorplot() for ggplot2 compatibility:
ggadjust_pvalue() added to adjust p-values produced by geom_pwc() on a ggplot (#522).gene_expressionggpubr.null_device, whose value should be a function that creates an appropriate null device. These include: cowplot::pdf_null_device, cowplot::png_null_device, cowplot::cairo_null_device and cowplot::agg_null_device. Default is cowplot::pdf_null_device. This is used in functions like as_ggplot(), which need to open a graphics device to render ggplot objects into grid graphics objects. This function is used to open a null device to avoid displaying an unnecessary blank page when calling ggarrange() or as_ggplot() (#306 and #158). The default option can be changed using, for example, options(ggpubr.null_device = cowplot::png_null_device).gadd(): Restoring back random state after setting seed when adding jittered points. To do so, the seed number is just passed to position_jitter() and position_jitterdodge(), which preserve the initial random state ( #177 and #349) .ggpubr now requires a version of ggrepel >= 0.9.2.9999, which now restores the initial random state after set.seed(). See https://github.com/slowkow/ggrepel/issues/228ggpubr now requires a version of cowplot >= 1.1.1ggtexttable(): doc updated with another example; text justification for individual cells/rows/columns (#335).ggpie(): setting the default of clip = "off" in coord_polar() so that ggpie() does not crop labels (#429)as_ggplot(): using null_device to avoid blank page #306 and #158ggarrange(): using null_device to avoid blank page #306 and #158df is a tibble.ggexport(): support added for graphics device svg (#469)ggpie() and ggdonutchart() now fully reacts to the option lab.font (#502)gather_() in both internal (.check_data()) and exported functions (compare_means()) (#513)stat_compare_means(): The dot-dot notation (..p.signif..) was deprecated in ggplot2 3.4.0; after_stat(p.signif) should be used; updated so that ..p.signif.. is automatically converted into after_stat() format without warning for backward compatibility.desc_statby() doc updated to clarify the difference between SD (standard deviation) and SE (standard error) (#492)geom_smooth() using formula 'y ~ x' is now turned off in ggscatter()(#488)ggtext(): fix warning "filter_() was deprecated in dplyr 0.7.0".ggqqplot(): the argument conf.int is taken into account now when specified (#524).ggqqplot(): Fixing the warning: "The following aesthetics were dropped during statistical transformation: sample" (#523)rstatix v >=0.7.1.999 for preserving factor class in emmeans_test() (#386)ggmaplot(): Suppressing ggmaplot warning: Unlabeled data points (too many overlaps). Consider increasing max.overlaps (#520)compare_means(): works now when the grouping variable levels contain the key words group2 or group1 (#450)ggparagraph() : fixing bug about minimum paragraph length (#408)ggexport(): the verbose argument is now considered when specified by user (#474)stat_anova_test(), stat_kruskal_test(), stat_welch_anova_test(), stat_friedman_test() and geom_pwc() added. These are flexible functions to add p-values onto ggplot with more options. The function geom_pwc() is for adding pairwise comparisons p-values to a ggplot; supported statistical methods include "wilcox_test", "t_test", "sign_test", "dunn_test", "emmeans_test", "tukey_hsd" and "games_howell_test".as_npc(), npc_to_data_coordinates() and get_coord().ggpubr_options() to display allowed global options in ggpubrggpubr.parse_aes. Logical indicating whether to parse aesthetic variable names. Default is TRUE. For example, if you want ggpubr to handle non-standard column names, like A-A, without parsing, then set this option to FALSE using options(ggpubr.parse_aes = FALSE).stat_conf_ellipse: ensure stat returns a data.frame for compatibility with ggplot2 v>=3.4.0stat_compare_means():
after_stat(p.signif) as the dot-dot notation (..p.signif..) was deprecated in ggplot2 3.4.0 (#509).ggdensity() and gghistogram(): dot-dot notation (..density.., ..count..) replaced by after_stat(density) and after_stat(count), respectively for compatibility with ggplot2 3.4.0.create_aes():
digits and table.font.size) added to ggsummarystats() for changing the summary table decimal place and text size (#341).stat_pvalue_manual() the argument hide.ns can be either a logical value (TRUE or FALSE) or a character value ("p" or "p.adj" for filtering out non significant by p-value or adjusted p-values).vjust = 0.5 (#301).Capital NS. is no longer displayed by stat_compare_means() (#171)ggshistogram() to make sure that it works when:
after_stat(),ggscatter() to make sure that:
ggpubr.parse_aes global option is set to FALSE (#229)ggtexttable() (#125, #129 and #283):
tab_cell_crossout(): cross out a table cell.tab_ncol(), tab_nrow(): returns, respectively, the number of columns and rows in a ggtexttable.tab_add_hline(): Creates horizontal lines or separators at the top or the bottom side of a given specified row.tab_add_vline(): Creates vertical lines or separators at the right or the left side of a given specified column.tab_add_border(), tbody_add_border(), thead_add_border(): Add borders to table; tbody is for table body and thead is for table head.tab_add_title() and tab_add_footnote() to add titles and footnotes (#243).create_aes() added to create aes mapping from a list. Makes programming easy with ggplot2 (#229).coord.flip added to support adding p-values onto horizontal ggplots (#179). When adding the p-values to a horizontal ggplot (generated using coord_flip()), you need to specify the option coord.flip = TRUE.median_hilow_() and median_q1q3() - added (@davidlorenz, #209):
median_hilow_(): computes the sample median and a selected pair of outer quantiles having equal tail areas. This function is a reformatted version of Hmisc::smedian.hilow(). The confidence limits are computed as follows: lower.limits = (1-ci)/2 percentiles; upper.limits = (1+ci)/2 percentiles. By default (ci = 0.95), the 2.5th and the 97.5th percentiles are used as the lower and the upper confidence limits, respectively. If you want to use the 25th and the 75th percentiles as the confidence limits, then specify ci = 0.5 or use the function median_q1q3().median_q1q3(): computes the sample median and, the 25th and 75th percentiles. Wrapper around the function median_hilow_() using ci = 0.5.get_breaks() added to easily create breaks for numeric axes. Can be used to increase the number of x and y ticks by specifying the option n. It's also possible to control axis breaks by specifying a step between ticks. For example, if by = 5, a tick mark is shown on every 5 (@Chitanda-Satou, #258).ggscatterhist() (@juliechevalier, #176):
ggscatterhist() is now a list of ggplots, containing the main scatter plot (sp) and the marginal plots (xplot and yplot), which can be customized by the end user using the standard ggplot verbsalternative supported in stat_cor() (#276).position in ggline() to make position "dodged" (#52).outlier.shape in ggboxplot(). Default is 19. To hide outlier, specify outlier.shape = NA. When jitter is added, then outliers will be automatically hidden.ggdotchart() using the option sorting = "none" (#115, #223).weight added in gghistogram() for creating a weighted histogram (#215)ggscaterhist() takes into account the argument position in margin.params when marginal plot is a histogram (#286).ggbarplot() enhanced to better handle the creation of dodged bar plots combined with jitter points (@aherholt, #176)bracket.shorten added in stat_pvalue_manual() and geom_bracket(). a small numeric value in [0-1] for shortening the with of bracket (#285).bracket.nudge.y added in stat_pvalue_manual() and geom_bracket(). Vertical adjustment to nudge brackets by. Useful to move up or move down the bracket. If positive value, brackets will be moved up; if negative value, brackets are moved down (#281).numeric.x.axis added in ggerrorplot(); logical value, If TRUE, x axis will be treated as numeric. Default is FALSE (#280).width is now considered in ggadd() for plotting error bars (#278).linetype in ggpaired().geom_exec() used in ggpaired() to add lines between paired points.ggmaplot() now supports two input formats (#198):
ggmaplot():
alpha for controlling point transparency/density (@apcamargo, #152).label.select to select specific genes to show on the plot (@apastore, #70)ggadd() the fill argument is considered for jitter points only when the point shape is in 21:25 (@atakanekiz, #148).parse added in ggscatter() and in ggtext(). If TRUE, the labels will be parsed into expressions and displayed as described in ?plotmath (#250).stroke supported in ggscatter() and in ggline(). Used only for shapes 21-24 to control the thickness of points border (@bioguy2018, #258).stat_cor() function code has been simplified. New arguments p.accuracy and r.accuracy added; a real value specifying the number of decimal places of precision for the p-value and the correlation coefficient, respectively. Default is NULL. Use (e.g.) 0.01 to show 2 decimal places of precision (@garthtarr, #186, @raedevan6, #114, #270).annotate_figure() manual updated to show how to use of superscript/subscript in the axis labels (#165).ggtextable() now supports further customization when theme is specified (#283).font.family is now correctly handled by ggscatter() (#149)ggpar() arguments are correctly applied using ggpie() (#277).ggscatter(): When conf.int = FALSE, fill color is set to "lightgray" for the regression line confidence band (@zhan6073, #111).gghistogram() supports the parameter yticks.by (@Chitanda-Satou, #258).ggsummarystats() to create a GGPlot with summary stats table under the plot ( #251).clean_table_theme() to clean the theme of a table, such as those created by ggsummarytable()ggbarplot() now supports stacked barplots with error bars (#245).vjsut in stat_compare_means() to move the text up or down relative to the bracket.type in geom_bracket() to specify label type. Can be "text" or "expression" (for parsing plotmath expression); #253.labeller to the function facet()position in get_legend() to specify legend positionlegend.grob in ggarrange() to specify a common legend you want to add onto the combined plot.cor.coef.name in the function stat_cor(). Can be one of "R" (pearson coef), "rho" (spearman coef) and "tau" (kendall coef). Uppercase and lowercase are allowed (@andhamel, #216).digits, r.digits, p.digits in the function stat_cor(). Integer indicating the number of decimal places (round) or significant digits (signif) to be used for the correlation coefficient and the p-value (@raedevan6, #216).compare_means() adapted to tidyr v>= 1.0.0 by specifying cols in the unnest() function (@Youguang, #216).stat_pvalue_manual() can now handle an rstatix test result containing only one group column.stat_central_tendency() to add central tendency measures (mean, median, mode) to density and histogram plotsstat_overlay_normal_density() to overlay normal density plot (with the same mean and SD) to the density distribution of 'x'.exact = FALSE is no longer used when computing correlation in stat_cor() (@tiagochst, #205)ggpie() keeps now the default order of labels (@WortJohn, #203)geom_bracket() for adding brackets with label annotation to a ggplot. Helpers for adding p-value or significance levels to a plot.compare_means() has been adapted to tidyr v1.0.0 (@jennybc, #196)geom_exec() now handles geom_bracket() argumentsvjust, hide.ns, step.increase, step.group.by, color and linetype added in stat_pvalue_manual()stat_pvalue_manual() can now guess automatically the significance label column.show.legend added to ggadd() and add_summary() functions.gghistogram(). Works now when the x variable is R keyword, such as var, mean, etc. (#192)ggline(), error bars now react automatically to grouping by line type (#191)step.increase added in stat_compare_means() to avoid overlap between brackets.stat_pvalue_manual() x axis variable is no longer automatically converted into factor. If your x variable is a factor, make sure that it is converted into factor.stat_pvalue_manual() can automatically handle the output of rstatix testsggbarplot() and ggviolin() now automatically create error bars by groups when users forget the option add.params = list(group = ) (#183).ggarrange() works when either ncol = 1 or nrow = 1 (@GegznaV, #141.compare_means() set automatically the option exact = FALSE. This is no longer the case (@stemicha, #141.stat_pvalue_manual() now supports dodged grouped plots (@emcnerny, #104).position is now handled by ggdotplot() (@Adam-JJJJJ, #178)label.sep argument works now in ggscatter() and stat_cor() (@sbbmu, #150)ggscatter() to avoid freezing when the add argument is incorrect (@atakanekiz, #135).The option ref.group was only considered when the grouping variable contains more than two levels. In that case, each level is compared against the specified reference group. Now, ref.group option is also considered in two samples mean comparisons (@OwenDonohoe, #118)
Now, ggqqplot() reacts to the argument conf.int.level (@vsluydts, #123)
Added error bar color is now inherited from the main plot (@JesseRop, #109)
bxp.errorbar added to ggboxplot() for adding error bars at the top of the box plots (@j3ypi, #105.stat_pvalue_manual() for adding p-values generated elsewhere (@achamess, #81, @grst, #65).alpha option added to ggviolin() @mtmatter, #77bracket.size added to stat_compare_means() @mtmatter, #43stat_cor() supports R^2 as an option @philament, #32position added in gghistogram(). Allowed values include "identity", "stack", "dodge".ci added in ggerrorplot() @abrar-alshaer, #94ggscatter() can remove the letter 'a' from the legend, when the argument show.legend.text = FALSE specified @atsyplenkov, #106.size option to ggscatter add.params is supported @retrogenomics, #94.ggdonutchart() added.Significance levels can be now customized and passed to stat_compare_means() (@jaison75, #45).
Editing pdf size is now supported in ggexport() (@JauntyJJS, #45).
ggscatterhist() the x variable was plotted two times, on both the plot x & y margins, instead of having, as expected, a) the x variable on the main plot x margin and 2) the y variable on the main plot y margin. This has been now fixed.ggdotchart() sorted automatically within groups when the color argument is specified, even when groups = NULL. This default behaviour has been now removed. Sorting within groups is performed only when the argument group is specified (@sfeds, #90).yticks.by and xticks.by work with NAs (@j3ypi, #89).New function ggballoonplot() added to visualize a contingency table.
ggdotchart() can be now used to plot multiple groups with position = position_dodge() (@ManuelSpinola, #45).
New function ggscatterhist() to create a scatter plot with marginal histograms, density plots and box plots.
New theme theme_pubclean(): a clean theme without axis lines, to direct more attention to the data.
New arguments in ggarrange() to customize plot labels (@G-Thomson, #41):
New argument method.args added to stat_compare_means(). A list of additional arguments used for the test method. For example one might use method.args = list(alternative = "greater") for wilcoxon test (@Nicktz, #41).
New argument symnum.args added to stat_compare_means(). A list of arguments to pass to the function symnum for symbolic number coding of p-values. For example, symnum.args <- list(cutpoints = c(0, 0.0001, 0.001, 0.01, 0.05, 1), symbols = c("****", "***", "**", "*", "ns"))
New functions table_cell_font() and table_cell_bg() to easily access and change the text font and the background of ggtexttable() cells (@ProbleMaker, #29).
New argument numeric.x.axis in ggline(). logical. If TRUE, x axis will be treated as numeric. Default is FALSE. (@mdphan, #35)
New argument lab.nb.digits in ggbarplot(). Integer indicating the number of decimal places (round) to be used (#28). Example: lab.nb.digits = 2.
New argument tip.length in stat_compare_means(). Numeric vector with the fraction of total height that the bar goes down to indicate the precise column. Default is 0.03. Can be of same length as the number of comparisons to adjust specifically the tip length of each comparison. For example tip.length = c(0.01, 0.03).
get_legend() returns NULL when the plot doesn't have legend.Now data argument are supported in stat_compare_means() when the option comparisons are specified (@emcnerny, #48)
Now compare_means() returns the same p-values as stat_compare_means() (@wydty, #15).
stat_compare_means() now reacts to label = "p.format" when comparisons specified (#28).
Now, the p.values are displayed correctly when ref.group is not the first group (@sehufnkjesktgna, #15).
In ggpar(), now legend.title can be either a character vector, e.g.: legend.title = "Species" or a list, legend.title = list(color = "Species", linetype = "Species", shape = "Species").
New argument ellipse.border.remove in ggscatter() to remove ellipse border lines.
ggscatter(mtcars, x = "mpg", y = "wt",
color = "cyl",
ellipse = TRUE, mean.point = TRUE,
ellipse.border.remove = TRUE)
In ggscatter(), the argument mean.point now reacts to fill color.
Support for text justification added in ggtexttable() (@cj-wilson, #15)
The function ggpie() can now display japanese texts. New argument font.family in ggpie() and in ggpar() (@tomochan001, #15).
Using time on x axis works know with ggline() and ggbarplot() (@jcpsantiago, #15).
stat_compare_means() now reacts to hide.ns properly.drawDetails.splitText() exported so that the function ggparagraph() works properly.ggbarplot(), now labels correspond to the true size of bars (@tdelhomme, #15).stat_compare_means() now keeps the default order of factor levels (@RoKant, #12).New helper functions:
gradient_color() and gradient_fill(): change gradient color and fill palettes.clean_theme(): remove axis lines, ticks, texts and titles.get_legend(): to extract the legend labels from a ggplot object.as_ggplot(): Transform the output of gridExtra::arrangeGrob() and gridExtra::grid.arrange() to an object of class ggplot.ggtexttable(): to draw a textual table.ggparagraph(): to draw a paragraph of text.annotate_figure() to annotate (arranged) ggplots.text_grob() to create easily a customized text graphical object.background_image() to add a background image to a ggplot.New theme function theme_transparent() to create a ggplot with transparent background.
gghistogram(), density curve and rug react to the fill color.ggarrange():
align to specify whether graphs in the grid should be horizontally ("h") or vertically ("v") aligned.legend to remove or specify the legend position when arranging multiple plots.common.legend to create a common unique legend for multiple plots.New functions:
ggarrange() to arrange multiple ggplots on the same page.ggexport() to export one or multiple ggplots to a file (pdf, eps, png, jpeg).ggpaired() to plot paired data.compare_means() to compare the means of two or multiple groups. Returns a data frame.stat_compare_means() to add p-values and significance levels to plots.stat_cor() to add correlation coefficients with p-values to a scatter plot.stat_stars() to add stars to a scatter plot.Now, the argument y can be a character vector of multiple variables to plot at once. This might be useful in genomic fields to plot the gene expression levels of multiple genes at once. see ggboxplot(), ggdotplot(), ggstripchart(), ggviolin(), ggbarplot() and ggline.
The argument x can be a vector of multiple variables in gghistogram(), ggdensity(), ggecdf() and ggqqplot().
New functions to edit ggplot graphical parameters:
font() to change the appearance of titles and labels.rotate_x_text() and rotate_y_text() to rotate x and y axis texts.rotate() to rotate a ggplot for creating horizontal plot.set_palette() or change_palette() to change a ggplot color palette.border() to add/change border lines around a ggplot.bgcolor() to change ggplot panel background color.rremove() to remove a specific component from a ggplot.grids() to add grid lines.xscale() and yscale() to change axis scale.New helper functions:
facet() added to create multi-panel plots (#5).add_summary() to add summary statistics.ggadd() to add summary statistics or a geometry onto a ggplot.New data set added: gene_citation
New arguments in ggpar(): x.text.angle and y.text.angle
New arguments in ggpubr functions, see ggboxplot(), ggdotplot(), ggstripchart(), ggviolin(), ggbarplot() and ggline:
combine added to combine multiple y variables on the same graph.merge to merge multiple y variables in the same plotting area.select to select which item to display.remove to remove a specific item from a plot.order to order plot items.label, font.label, label.select, repel, label.rectangle to add and customize labelsfacet.by, panel.labs and short.panel.labs: support for faceting and customization of plot panelsNew argument grouping.vars in ggtext(). Grouping variables to sort the data by, when the user wants to display the top n up/down labels.
New arguments in theme_pubr():
palette can be also a numeric vector of length(groups); in this case a basic color palette is created using the function grDevices::palette().Now, ggpar() reacts to palette when length(palette) = 1 and palette is a color name #3.
ggmaplot() now handles situations, where there is only upregulated, or downregulated genes.
get_palette() to generate a palette of k colors from ggsci palettes, RColorBrewer palettes and custom color palettes. Useful to extend RColorBrewer and ggsci to support more colors.ggpar() function can handle a list of ggplots.right.show.legend.text in the ggscatter() function. Use show.legend.text = FALSE to hide text in the legend.title, submain, subtitle, caption, font.submain, font.subtitle, font.caption in the ggpar() function.font.family in ggscatter().ggdensity (gghistogram) are now shown when data have NA values @chunkaowang, #1ggtext() for textual annotation.ggscatter(). A logical value. If TRUE, a star plot is generated.geom_exec(). A helper function used by ggpubr functions to execute any geom_xx functions in ggplot2. Useful only when you want to call a geom_xx function without worrying about the arguments to put in ggplot2::aes().ggbarplot().
theme_classic2() added. Classic theme with axis lines.ggboxplot(), ggviolin(), ggdotplot(), ggstripchart(), gghistogram(), ggdensity(), ggecdf() and ggqqplot() can now handle one single numeric vector.# Example
ggboxplot(iris$Sepal.Length)
gghistogram(), when add_density = TRUE, y scale remains = "..count..".