Package 'leo.basic'

Enrichment helpers

Description

Utilities to format gene lists and collapse IDs for KEGG/Reactome GSEA/ORA.

• You can download gmt file from https://maayanlab.cloud/Enrichr/#libraries

• Tutorial: https://yulab-smu.top/biomedical-knowledge-mining-book/faq.html#genelist

Build a named numeric vector sorted decreasingly

Format enrichment gene string like "A/B/C" to vector

Map SYMBOL-named vector to ENTREZID and keep 1 score/gene by slice_max(|score|)

Usage

format_geneList(df)

format_gene_str2vec(x, sep = "/", unique_only = TRUE)

prep_GSEA_geneList(geneList_sym)

Arguments

df	A data frame with two column: Column 1: 1st column is Gene ID Column 2: 2nd column is Gene Statistics (e.g., fold change or other numerical variable)
x	A character vector or a list of character vectors
sep	Separator, default "/"
unique_only	Return unique identifiers, default TRUE
geneList_sym	Named numeric vector; names are SYMBOL; values are scores

Details

Enrichment helpers

Value

A named numeric vector (descending)

A character vector or a list of character vectors

Named numeric vector whose names are unique ENTREZID (descending)

Functions

• format_geneList: make a named numeric vector sorted decreasingly

• format_gene_str2vec: split core_enrichment string into gene vector

• prep_GSEA_geneList: map names to ENTREZID and keep 1 score/gene by |score| max

See Also

clusterProfiler, ReactomePA

Examples

# ego4@result$core_enrichment[1] -> "IGLV3-19/IGHM/IGLV7-43/HLA-DPA1/IGHV3-7"
# format_gene_str2vec(ego4@result$core_enrichment[1])
## Not run: 
data(geneList, package = "DOSE")
gl_sym <- stats::setNames(as.numeric(geneList), names(geneList)) # demo only
gl_entrez <- prep_GSEA_geneList(gl_sym)

## End(Not run)

---

Enrichment wrappers (individual)

Description

ORA/GSEA wrappers for GO, KEGG, KEGG Module, and Reactome.

Usage

ORA_GO(gene, simplify = TRUE)

GSEA_GO(geneList, simplify = TRUE)

ORA_KEGG(gene, input = "SYMBOL")

GSEA_KEGG(geneList, input = "SYMBOL")

ORA_MKEGG(gene, input = "SYMBOL")

GSEA_MKEGG(geneList, input = "SYMBOL")

ORA_Reactome(gene, input = "SYMBOL")

GSEA_Reactome(geneList, input = "SYMBOL")

Arguments

gene	SYMBOL or ENTREZID vector depending on input
simplify	Use clusterProfiler::simplify for GO terms
geneList	Named numeric vector (SYMBOL or ENTREZID per input)
input	"SYMBOL" or "ENTREZID"

Details

Enrichment wrappers (individual)

Value

enrichResult

gseaResult

enrichResult

gseaResult

enrichResult

gseaResult

enrichResult

gseaResult

Functions

• ORA_GO, GSEA_GO

• ORA_KEGG, GSEA_KEGG

• ORA_MKEGG, GSEA_MKEGG

• ORA_Reactome, GSEA_Reactome

See Also

leo_enrich

Examples

## Not run: 
# KEGG vis notes:
# https://yulab-smu.top/biomedical-knowledge-mining-book/clusterprofiler-kegg.html
library("pathview")
kk <- gseKEGG(geneList     = geneList,
              keyType      = 'SYMBOL',
              organism     = 'hsa',
              minGSSize    = 120,
              pvalueCutoff = 0.05,
              verbose      = FALSE)
browseKEGG(kk, 'hsa04110')
# or
hsa04110 <- pathview(
  gene.data  = geneList,
  pathway.id = "hsa04110",
  species    = "hsa",
  limit      = list(gene = max(abs(geneList)), cpd = 1)
)

## End(Not run)

---

Install LEO package dependencies

Description

Install dependency sets for leo.basic or other leo.* packages universe using pak.

Usage

install_deps(leo.pak = "leo.basic", ncpus = 4L, upgrade = FALSE)

Arguments

leo.pak	Character scalar. LEO package name. Default "leo.basic".
ncpus	Integer. Number of CPU cores for parallel compilation. Default 4.
upgrade	Logical. Whether to upgrade already installed packages. Default FALSE.

Value

Invisibly returns the package specs passed to pak::pkg_install().

Examples

## Not run: 
# 1) Install pak
install.packages("pak", repos = "https://cloud.r-project.org")

# 2) Install leo.basic first
pak::pkg_install("laleoarrow/leo.basic")

# 3) Install leo.ukb (deps first, then package)
leo.basic::install_deps("leo.ukb", ncpus = 8)
pak::pkg_install("laleoarrow/leo.ukb")

# 4) Install leo.gwas (deps first, then package)
leo.basic::install_deps("leo.gwas", ncpus = 8)
pak::pkg_install("laleoarrow/leo.gwas")

## End(Not run)

---

Install local package

Description

This function is writen as I often forget what to code when I want to install a package from local source. This now uses pak::pkg_install() for a consistent installation backend. Anyway, it is always good to simplify the process.

Usage

install_local(path)

Arguments

path	The path to the local package source.

---

Generate a discrete color palette (expanded from a base panel)

Description

Return n distinct colors or a named vector for levels. Uses a fixed base panel and expands smoothly if more colors are needed.

Usage

leo_discrete_color(levels = NULL, n = NULL, base_panel = NULL)

Arguments

levels	character; category names. If provided, output is named with levels. If NULL, use n.
n	integer; number of colors to return when levels is NULL.
base_panel	named character; optional seed palette (hex). Defaults to an internal panel.

Value

character vector of hex colors; named if levels is provided.

Examples

leo_discrete_color(n = 8)
leo_discrete_color(levels = c("Brain","Liver","Heart"))

---

Enrichment analysis (Integrated)

Description

Enrichment analysis for GO/KEGG/MKEGG/Reactome using ORA or GSEA; iterate method first, then background.

Usage

leo_enrich(
  gene,
  geneList,
  simplify = TRUE,
  input = "SYMBOL",
  method = c("ORA", "GSEA"),
  background = c("GO", "KEGG", "MKEGG", "Reactome")
)

Arguments

gene	SYMBOL vector for ORA
geneList	Named numeric vector for GSEA (names match input)
simplify	Whether to simplify GO results
input	"SYMBOL" or "ENTREZID" (for KEGG/MKEGG/Reactome)
method	c("ORA","GSEA"), supports multiple
background	c("GO","KEGG","MKEGG","Reactome"), supports multiple

Value

A list of enrichment results per method-background combo

Examples

## Not run: 
# We here use SYMBOL in the first place
library(org.Hs.eg.db)
data(geneList, package="DOSE")
gene.df <- tibble(logFC=geneList, ID=names(geneList))
gene.df.map <- bitr(gene.df$ID, fromType = "ENTREZID", # optional
                    toType = c("SYMBOL"),
                    OrgDb = org.Hs.eg.db)
gene.df <- gene.df %>% left_join(gene.df.map, by = c("ID" = "ENTREZID")) %>% drop_na()
gene.df <- gene.df %>% arrange(desc(logFC))
gene <- gene.df$SYMBOL[1:100] # for ORA
geneList <- setNames(gene.df$logFC, gene.df$SYMBOL)  # for GSEA

## End(Not run)

---

Log messages with time stamps

Description

Logs messages with time stamps The messages are styled using the cli package for enhanced readability. This function can not deal with function in the cli package.

Usage

leo_log(..., level = "info", verbose = TRUE)

Arguments

...	The message string to log, which will be pasted together.
level	The log level. Options are "info", "success", "warning", and "danger".
verbose	in case you want to turn the info log off, set it to FALSE.

Value

No return value. Outputs a formatted log message with a timestamp.

Examples

n1 <- 10; n2 <- 20
leo_log("Processing the", n1, "and", n2, "files.")
leo_log("Task completed successfully!", level = "success")
leo_log("Potential issue detected.", level = "warning")
leo_log("Error occurred during processing!", level = "danger")

---

Give messages with my color

Description

Give messages with my color

Usage

leo_message(..., color = "31", return = FALSE)

Arguments

...	The messages you wanna messgae, which will be pasted together
color	Str. Preferred color. Default is yellow. Options are "31" (red), "32" (green), "34" (blue), "95" (light purple)...
return	Logical. If TRUE, returns the formatted string. If FALSE (Default), prints directly.

Examples

leo_message("This is a red message", color = "31")
leo_message("This is a green message", "\nhaha", color = "32")
leo_message("This is a blue ", "message", color = "34")
leo_message("This is a light purple message", color = "95")
leo_message("Welcome to use the LEO package!")

---

Generate a custom colour palette

Description

Convenience wrapper around colorRampPalette() that follows my-preferred coding conventions.

Usage

leo_more_color(color_set, expected_num)

Arguments

color_set	Character vector. Base colours to interpolate between (e.g. c("#1f77b4", "#ff7f0e", "#2ca02c")).
expected_num	Integer (scalar). Number of colours you want to generate.

Value

A character vector of hexadecimal colour codes of length expected_num.

Examples

# 10-colour palette smoothly blended from three anchor colours
leo_more_color(c("#1f77b4", "#ff7f0e", "#2ca02c"), 10)

---

leo_permutation_test: flexible one-way permutation / rank-based testing

Description

Compare one or multiple numeric features between two groups using permutation-based tests from coin. Two usage modes:

1. Vector mode - supply a numeric vector data and a grouping vector class_vec.

2. Data-frame mode - supply a data.frame data plus the column name holding the groups (class_col) and, optionally, the feature columns (feature_cols).

Workflow per feature:

• Normality test - Shapiro-Wilk for $n <= 5000$, Anderson-Darling for $n > 5000$.

• Variance homogeneity - F-test.

• If both criteria hold, run coin::oneway_test; otherwise run coin::wilcox_test.

Usage

leo_permutation_test(
  data,
  class_vec = NULL,
  class_col = "Class",
  feature_cols = NULL,
  set_seed = 725
)

Arguments

data	Numeric vector or data.frame.
class_vec	Grouping vector when data is numeric (length must match data); accepted as 0/1 or any two-level factor/character.
class_col	Character. Column name of grouping variable in a data.frame. Default "Class".
feature_cols	Character vector of feature columns in the data.frame. If NULL (default), all numeric columns except class_col are used.
set_seed	Integer random seed (passed to set.seed). Default 725.

Value

A tibble with three columns:

Feature

Feature name

P

P-value

Type

Test used: "oneway" or "wilcox"

Examples

## vector mode --------------------------------------------------------------
set.seed(1)
feat <- rnorm(100); cls <- rep(c(0, 1), each = 50)
leo_permutation_test(feat, class_vec = cls)

## data-frame mode: specify feature columns ---------------------------------
df <- data.frame(Class = cls,
                 F1 = feat,
                 F2 = rnorm(100, 0.3),
                 F3 = rnorm(100, 1))
leo_permutation_test(df, class_col = "Class", feature_cols = c("F1", "F2", "F3"))

## data-frame mode: all numeric features except Class -----------------------
leo_permutation_test(df, class_col = "Class")

## large-n example triggers AD test -----------------------------------------
big_feat <- rnorm(6000); big_grp <- rep(c(0, 1), each = 3000)
leo_permutation_test(big_feat, class_vec = big_grp)

---

Stop execution with formatted message

Description

Combines leo_log(..., level = "danger") with stop() for consistent error handling. Prints a timestamped red message before stopping execution.

Usage

leo_stop(..., call. = FALSE)

Arguments

...	Messages to combine (passed to leo_log)
call.	Logical; if FALSE (default), the call is not included in error message

Value

This function never returns; it stops execution with an error.

Examples

## Not run: 
if (!is.data.frame(df)) leo_stop("df must be a data.frame.")

## End(Not run)

---

Custom ggplot2 theme

Description

I often forget my favorite ggplot theme settings, so I write this to reminds myself.

Usage

leo_theme(
  plot_title_size = 14,
  legend_title_size = 10,
  legend_text_size = 9,
  axis_title_size = 10,
  axis_text_size = 8,
  type = c("console", "object")
)

Arguments

plot_title_size	Numeric; title font size (default 14).
legend_title_size	Numeric; legend title font size (default 10).
legend_text_size	Numeric; legend text size (default 9).
axis_title_size	Numeric; axis title size (default 10).
axis_text_size	Numeric; axis text size (default 8).
type	Character; "object" to return theme, or "console" (default) to print copyable code.

Examples

library(ggplot2)
# Add to a plot
ggplot(mtcars, aes(wt, mpg)) + geom_point() +
  leo_theme()
# Print theme code
leo_theme(type = "console")

---

Intelligent Elapsed-Time Formatter

Description

Compute and format the time difference between now and start_time, choosing the most appropriate unit automatically:

• ⁠< 1 s⁠ -> milliseconds (ms)

• ⁠< 60 s⁠ -> seconds (sec)

• ⁠< 1 h⁠ -> minutes (min)

• ⁠< 1 day⁠ -> hours (hr)

• ⁠< 1 week⁠ -> days (day)

• ⁠< 1 month⁠-> weeks (wk)

• ⁠< 1 year⁠ -> months (mon)

• otherwise -> years (yr)

Usage

leo_time_elapsed(start_time, digits = 2, return = FALSE)

Arguments

start_time	POSIXct. output from Sys.time().
digits	Integer. digits to keep (defaut: 2).
return	Logical. if TRUE, only return the calculated results.

Value

If return = FALSE (default), prints the elapsed time in a formatted string

Examples

t0 <- Sys.time() - 0.123
leo_time_elapsed(t0) # -> [12:34:56] Elapsed 125 ms

t1 <- Sys.time() - 3.5
leo_time_elapsed(t1) # -> [12:35:00] Elapsed 3.5 sec

# Capture string without printing
t2 <- Sys.time() - 61
elapsed_str <- leo_time_elapsed(t2, return = TRUE)
print(elapsed_str) # "1.08 min"

---

Plot numeric values by group with jitter / beeswarm, color / shape / size mapping

Description

Draw a grouped scatter of numeric values with optional jitter / beeswarm arrangement, color / shape / size encoding, plus optional mean indicator.

Usage

plot_group_numbers(
  df,
  group,
  number,
  color_col = NULL,
  color_rule = NULL,
  shape_col = NULL,
  shape_rule = NULL,
  size_col = NULL,
  size_rule = NULL,
  jitter = c("no", "yes", "bee"),
  x_axis_pos = c("default", "zero"),
  mean_type = c("none", "point", "line"),
  legend = TRUE
)

Arguments

df	data.frame.
group	column name (chr) mapped to x-axis categories.
number	column name (chr) mapped to y-axis numeric values.
color_col	column name (chr) for point color (skip color_rule).
color_rule	function(df) -> color vector; ignored if color_col given.
shape_col	column name (chr) for point shape (skip shape_rule).
shape_rule	function(df) -> shape vector; ignored if shape_col given.
size_col	column name (chr) for point size (skip size_rule).
size_rule	function(df) -> size vector; ignored if size_col given.
jitter	one of "no", "yes", "bee".
x_axis_pos	"default" or "zero" (draw baseline at y = 0).
mean_type	"none", "point", "line".
legend	Logical. Whether to show legends for colour/shape/size mappings. Default TRUE. Set FALSE to hide all legends.

Value

ggplot object.

Note

Shape tips (ggplot shape codes): Run ⁠plot(0:25, pch = 0:25); text(0:25, labels = 0:25, pos = 3)⁠ to preview every shape code.

Code	Marker	Visual meaning	Best use
24	▲	solid triangle pointing up	filled point with border; use stroke for outline width
25	▼	solid triangle pointing down	filled point with border; use stroke for outline width
22	■	solid square	filled point with border; use stroke for outline width
0	□	open square	outline-only marker
2	△	open triangle pointing up	outline-only marker
6	▽	open triangle pointing down	outline-only marker
3	+	plus sign	thin marker for dense overlaps
4	x	x/cross sign	thin marker for dense overlaps
1	○	open circle	thin marker for dense overlaps

Shapes 21-25 accept separate border (colour) and interior (fill). For filled point with black outline: shape = 21, colour = "black", fill = "<fill>", tweak stroke. Provide *_col or *_rule, or leave both NULL. Defaults: colour sig/pos=red, sig/neg=blue, else grey; shape 16; size 2.

Examples

set.seed(1)
df <- data.frame(
  var        = paste0("gene", 1:60),
  group      = sample(c("Tcell","Bcell","NK"), 60, TRUE),
  aver_logFC = rnorm(60, 0, 2),
  p_val      = runif(60),
  my_shape   = sample(c(24,25,22), 60, TRUE),
  my_size    = sample(c(2,3,4), 60, TRUE)
)
plot_group_numbers(df,
                   group      = "group",
                   number     = "aver_logFC",
                   shape_col  = "my_shape",
                   size_col   = "my_size",
                   jitter     = "bee",
                   x_axis_pos = "zero",
                   mean_type  = "point")

---

Plot pie/ring via ggpie with install hint

Description

Plot pie/ring via ggpie with install hint

Usage

plot_pie(
  x,
  colors = NULL,
  color_alpha = 1,
  type = c("num", "ratio"),
  ring_ratio = 1,
  border = TRUE,
  annotation_type = c("in", "out"),
  label_type = c("horizon", "circle", "none"),
  label_size = 4
)

Arguments

x	Numeric vector: values (type='num') or ratios (type='ratio').
colors	Vector of fill colors; NULL uses package palette.
color_alpha	Alpha for colors in 0-1.
type	"num" or "ratio" for the meaning of x.
ring_ratio	Visible ring thickness in 0-1; 1=full pie (no hole).
border	Logical; draw slice borders or not (default TRUE).
annotation_type	"in" or "out" for label position.
label_type	"horizon" or "circle" for label style. Or "none" for no labels.
label_size	Label size.

Value

ggplot object.

Examples

plot_pie(
  c(A = 30, B = 20, C = 50),
  type = "num", ring_ratio = 1,
  annotation_type = "in"
)
plot_pie(
  c(A = 0.2, B = 0.3, C = 0.5),
  type = "ratio", ring_ratio = 0.6,
  annotation_type = "out",
  label_type = "horizon",
  colors = c("#66c2a5", "#fc8d62", "#8da0cb")
)
plot_pie(
  c(A = 30, B = 20, C = 50),
  ring_ratio = .6, border = FALSE,
  color_alpha = 0.8, type = "num",
  label_type = "horizon",
  annotation_type = "out"
)

---

Prism-style lollipop plot (grouped or single series)

Description

Create a lollipop plot with Prism-style x-axis brackets, suitable for grouped or single-series data.

Usage

plot_prism_lollipop(
  df,
  x_var,
  y_var,
  group_var = NULL,
  color_palette = NULL,
  y_label = "Value",
  y_as_percent = FALSE,
  plot_title = NULL,
  segment_width = 1,
  point_size = 4,
  point_stroke = 1
)

Arguments

df	Data frame containing the data.
x_var	Bare column name for the x-axis (categorical variable).
y_var	Bare column name for the y-axis (numeric variable).
group_var	Optional bare column name for grouping; if NULL, a single series is plotted.
color_palette	Named vector of colors for each group; for single-series, first color is used.
y_label	Label for the y-axis.
y_as_percent	Logical; if TRUE, scales y_var by 100 and formats labels as percentages.
plot_title	Optional plot title.
segment_width	Numeric; line width of the lollipop segments. Default is 1.
point_size	Numeric; size of the lollipop points. Default is 4.
point_stroke	Numeric; stroke width of point borders. Default is 1.

Value

A ggplot2 object.

Examples

# Single-series example
df_single <- data.frame(Stage = LETTERS[1:5], Rate = c(0.12, 0.14, 0.2, 0.17, 0.22))
plot_prism_lollipop(
  df_single, Stage, Rate,
  color_palette = c("#0072B2"),
  y_label = "Conversion Rate (%)",
  y_as_percent = TRUE,
  plot_title = "Single Series Lollipop Plot"
)

# Grouped example
df_grouped <- data.frame(Category = rep(LETTERS[1:5], 2), Score = runif(10, 0.3, 0.6),
                         Group = rep(c("G1", "G2"), each = 5))
pal <- c("G1" = "#E69F00", "G2" = "#56B4E9")
plot_prism_lollipop(
  df_grouped, Category, Score, Group,
  color_palette = pal,
  y_label = "Score (%)",
  y_as_percent = TRUE,
  plot_title = "Grouped Lollipop Plot"
)

---

Put legend inside ggplot/patchwork at (0.8, 0.8) npc

Description

Works for a single ggplot, a list of ggplots, or a patchwork object.

Usage

put_legend_inside(p, x = 0.8, y = 0.8)

Arguments

p	ggplot object, list of ggplot objects, or patchwork
x, y	npc coordinates in 0-1, default 0.8, 0.8

Value

Same type as input with legend repositioned

---

Rasterize point-like layers in a ggplot object

Description

Rasterize point-like layers in a ggplot object

Usage

rasterize_layers(
  plot,
  dpi = 300,
  layers = c("Point", "Jitter", "Line", "Segment", "Linerange", "EdgeSegment")
)

Arguments

plot	A ggplot object.
dpi	Integer DPI for rasterization (default 300).
layers	Character vector of layer types to rasterize, e.g. c("Point", "Jitter", "Line", "Segment").

Value

ggplot object with target layers rasterized.

Examples

library(ggplot2); library(ggrastr)
plot <- ggplot(diamonds, aes(carat, price, colour = cut)) + geom_point()
rasterize_layers(plot, dpi = 100)

---

Recode a Vector Using a Two-Column Mapping Table

Description

This function takes a vector x and a mapping data frame map_df that has two columns: one with the original values and one with the replacement values. Internally it builds a named vector via tibble::deframe(), then applies dplyr::recode() with splicing (!!!) to perform the recoding.

Usage

recode_by_map(x, map_df, from, to)

Arguments

x	A vector whose elements you want to recode.
map_df	A data frame or tibble containing exactly two columns: one for the old values, one for the new values.
from	A string giving the name of the column in map_df that contains the original values.
to	A string giving the name of the column in map_df that contains the replacement values.

Details

Build a 2-col mapping df, use deframe() then recode(col, !!!name_vector).

Value

A vector (same type as x) with all values matching map_df[[from]] replaced by the corresponding map_df[[to]]; all other values in x are left unchanged.

See Also

recode, deframe

Examples

library(dplyr)
map_df <- tibble::tibble(
  old = c("ENSG1", "ENSG2"),
  new = c("TP53",  "BRCA1")
)
recode_by_map(c("ENSG1","foo","ENSG2"), map_df, "old", "new")
#> [1] "TP53" "foo"  "BRCA1"

---

Set or unset proxy

Description

Quickly enable or disable HTTP/HTTPS (and optionally SOCKS5) proxy for the current R session.

Usage

set_proxy(enable = TRUE, host = "127.0.0.1", port = 7897, socks = TRUE)

Arguments

enable	TRUE to enable, FALSE to disable.
host	Proxy host address.
port	Proxy port number.
socks	Whether to also set SOCKS5 proxy.

Examples

# Enable proxy (HTTP/HTTPS + SOCKS5)
set_proxy(TRUE, host = "127.0.0.1", port = 7897, socks = TRUE)

# Test proxy
system("curl cip.cc")

# Disable proxy
set_proxy(FALSE)

---

View DataFrame with VisiData from Terminal

Description

vd() writes df to a temporary TSV file, then opens it with VisiData. This keeps the R side simple while letting VisiData read the table lazily from disk, which is much more usable for large data frames than printing them in the console. Non-default row names are preserved as the first column. A compact terminal summary is printed before launch, including dimensions, row-name handling, temp-file size, and writer backend.

Usage

vd(df)

Arguments

df	Data frame, tibble, or matrix to view.

Value

Invisibly returns df.

Examples

## Not run: 
data(mtcars)
vd(mtcars)
mtcars %>% vd()

## End(Not run)

---