SMR Miami Plot

Miami.r draws a Miami plot for comparing two SMR result sets. The first SMR file is displayed above the chromosome axis and the second file below it.

The script is intended for direct use with source() and is not an R package.

Requirements

install.packages(c("data.table", "ggplot2"))

Load the functions with:

source("miami.r")

Input format

Each input can be:

• a path to an SMR result file;
• a data.frame; or
• a data.table.

The default column names are:

Column	Description
probeID	Probe identifier used to compare signals between the two files
ProbeChr	Chromosome
Gene	Gene symbol used for labels
Probe_bp	Probe genomic position in base pairs
p_SMR	SMR p-value
p_HEIDI	HEIDI p-value

Column names can be changed in ReadMiamiData(). A single name is applied to both inputs; two names can be supplied when the files use different columns.

MiamiData <- ReadMiamiData(
    smr1 = "result_1.smr",
    smr2 = "result_2.smr",
    p.col = c("p_SMR", "SMR_P"),
    gene.col = c("Gene", "Symbol")
)

Quick start

source("miami.r")

MiamiData <- ReadMiamiData(
    smr1 = "result_1.smr",
    smr2 = "result_2.smr",
    heidi.threshold = 0.01,
    smr.alpha = 0.05
)

p <- MiamiPlot(
    data = MiamiData,
    plot.name = c("transSMR", "SMR (mismatch)"),
    highlight = TRUE,
    highlight.text = TRUE
)

print(p)

ggplot2::ggsave(
    "Miami_plot.pdf",
    p,
    width = 14,
    height = 8
)

Signal definition

The two input files are processed separately.

1. p_SMR is Bonferroni-adjusted using the number of unique probes in the corresponding file.
2. An SMR signal passes when p_SMR_adj < smr.alpha.
3. A signal passes HEIDI when p_HEIDI > heidi.threshold.
4. A top-specific signal is a probe that passes both tests in the first file but is absent from the probes passing both tests in the second file.

The plotting styles are:

Result	Plot style
Fails SMR	Chromosome-coloured background point
Passes SMR but fails HEIDI	Chromosome-coloured background point
Passes both SMR and HEIDI	Solid signal.col point
Top-specific signal	highlight.col point

ReadMiamiData()

This function reads, standardizes, compares, and prepares the two SMR result sets.

MiamiData <- ReadMiamiData(
    smr1,
    smr2,
    probe.col = "probeID",
    chr.col = "ProbeChr",
    gene.col = "Gene",
    pos.col = "Probe_bp",
    p.col = "p_SMR",
    heidi.col = "p_HEIDI",
    heidi.threshold = 0.01,
    smr.alpha = 0.05,
    mid.gap = NULL,
    p.min = 1e-300
)

Argument	Default	Description
smr1	required	First SMR file or data object; plotted on top
smr2	required	Second SMR file or data object; plotted on bottom
probe.col	"probeID"	Probe ID column; one name or two names
chr.col	"ProbeChr"	Chromosome column; one name or two names
gene.col	"Gene"	Gene column; use NULL to label with probe IDs
pos.col	"Probe_bp"	Genomic-position column
p.col	"p_SMR"	SMR p-value column
heidi.col	"p_HEIDI"	HEIDI p-value column; use NULL to disable HEIDI filtering
heidi.threshold	0.01	HEIDI passes when p_HEIDI is greater than this value
smr.alpha	0.05	Significance level applied to Bonferroni-adjusted SMR p-values
mid.gap	NULL	Central half-gap; NULL automatically uses 10% of maximum -log10(p_SMR)
p.min	1e-300	Lower p-value bound used before -log10()

The returned MiamiData object contains the standardized data, adjusted p-values, signal classifications, chromosome coordinates, thresholds, and axis positions required by MiamiPlot().

MiamiPlot()

This function returns a ggplot object and does not automatically write a file.

p <- MiamiPlot(
    data = MiamiData,
    plot.name = c("Top name", "Bottom name")
)

Required arguments

Argument	Description
data	Object returned by ReadMiamiData()
plot.name	Two labels corresponding to the top and bottom plots

Background and signal points

Argument	Default	Description
col	c("#BDBDBD", "#737373")	Background colours cycled across chromosomes
cex	0.9	Background-point size
pch	16	Background-point shape
alpha	0.65	Background-point transparency
signal.col	"#222222"	Colour for signals passing both SMR and HEIDI
signal.cex	2	Size of signals passing both tests

Highlighting

Argument	Default	Description
highlight	TRUE	TRUE selects automatic top-specific signals; a vector/list selects genes or probes manually; FALSE disables highlighting
highlight.by	"Gene"	Match manual values by "Gene", "probeID", or "both"
highlight.col	"#D73027"	Highlight-point colour
highlight.cex	2	Highlight-point size
highlight.pch	16	Highlight-point shape
highlight.text	TRUE	Show labels; FALSE keeps highlighted points without labels or leader lines
highlight.text.cex	1	Highlight-label size
highlight.text.col	"black"	Highlight-label colour
highlight.text.font	3	Label font: 1 plain, 2 bold, 3 italic, 4 bold italic
max.labels	30	Maximum number of labels, retaining the lowest p-values first; use Inf for all

Highlight labels use the original two-part leader-line layout: a vertical line rises from the point and then bends towards a vertical gene label above the plot. Label positions are ordered and spaced automatically along the genomic axis.

Thresholds, names, and text

Argument	Default	Description
threshold.col	"#D73027"	Threshold-line and threshold-text colour
threshold.lty	"dashed"	Threshold-line type
threshold.lwd	0.6	Threshold-line width
threshold.text	TRUE	Display the Bonferroni threshold values
plot.name.col	c("#74BF74", "#FFAA60")	Top and bottom plot-name colours
plot.name.cex	1.3	Plot-name size
chr.cex	1.2	Chromosome-label size
axis.cex	1.2	Y-axis tick-label and threshold-text size
lab.cex	1.2	X/Y-axis title size
title	NULL	Optional plot title
xlab	"Chromosome"	X-axis title
ylab	-log10(P_SMR) expression	Y-axis title

Highlight examples

Automatic top-specific signals with gene labels

p <- MiamiPlot(
    data = MiamiData,
    plot.name = c("transSMR", "SMR (mismatch)"),
    highlight = TRUE,
    highlight.text = TRUE
)

Highlight points without gene labels

p <- MiamiPlot(
    data = MiamiData,
    plot.name = c("transSMR", "SMR (mismatch)"),
    highlight = TRUE,
    highlight.text = FALSE
)

Manually specified genes

p <- MiamiPlot(
    data = MiamiData,
    plot.name = c("transSMR", "SMR (mismatch)"),
    highlight = list(Gene = c("GENE1", "GENE2", "GENE3")),
    highlight.by = "Gene",
    highlight.text = TRUE
)

Manual highlights are restricted to the first/top result set.

Appearance example

p <- MiamiPlot(
    data = MiamiData,
    plot.name = c("Method 1", "Method 2"),
    col = c("#BDBDBD", "#737373"),
    signal.col = "#222222",
    highlight.col = "#D73027",
    highlight.cex = 2.5,
    highlight.text.cex = 1.1,
    plot.name.cex = 1.4,
    chr.cex = 1.3,
    axis.cex = 1.3,
    lab.cex = 1.3
)

Batch plotting

run_interval_miami.R is a minimal example for matching and plotting multiple SMR file pairs:

Rscript run_interval_miami.R

It matches:

*_eQTLeas.smr  -> top plot
*_eQTLeur.smr  -> bottom plot

and saves one PDF per matched trait.

Reproducible simulated example

Generate two 13,000-gene SMR result sets and the example figure with:

Rscript simulate_smr_example.R

The generated files are written to simulation_output/.