QuantBayes Software Release

This directory contains the standalone QuantBayes binaries for macOS (Intel x86_64 and Apple silicon arm64) and Linux x86 64, along with the R package source release. The instructions below guide you through unpacking, verifying checksums, running the model, and installing the R package. The R package manual can be read here: Open the QuantBayes R manual. This software is released under MIT license; citation is appreciated.

Contents

Verify checksums

From inside software_release:

shasum -a 256 -c quantbayes_v1.0.0_linux_x86_64.tar.gz.sha256
shasum -a 256 -c quantbayes_v1.0.0_macos_x86_64.tar.gz.sha256
shasum -a 256 -c quantbayes_v1.0.0_macos_universal.tar.gz.sha256
shasum -a 256 -c quantbayes_v0.1.0_R.tar.gz.sha256

Option 1. Unpack and run the QuantBayes binary software

1. Extract

Linux:

tar -xzf quantbayes_v1.0.0_linux_x86_64.tar.gz
cd quantbayes

macOS:

The universal macOS binary runs natively on both Intel x86_64 based Macs and Apple silicon arm64 systems.

tar -xzf quantbayes_v1.0.0_macos_universal.tar.gz
cd quantbayes

If macOS blocks the binary on first run (only required for system install):

xattr -d com.apple.quarantine quantbayes

Each extracted directory contains:

quantbayes       (binary)
quantbayes.1     (manual page)
LICENSE
README.md
example_data/

2. Run QuantBayes

Human readable report:

./quantbayes example_data/test_matrix_01.txt --report --out sample1

JSON output:

./quantbayes example_data/test_matrix_01.txt --json

Most basic usage example:

./quantbayes example_data/test_matrix_01.txt

Optional add to PATH

mkdir -p $HOME/.local/bin
cp quantbayes $HOME/.local/bin/
export PATH="$HOME/.local/bin:$PATH"

Option 2. Use the R package which implements the same algorithm

Open the QuantBayes R manual.

The R package provided in this release is:

quantbayes_v0.1.0_R.tar.gz

Simply test with provided script in R studio:

R_example.R

Install with:

install.packages("quantbayes_v0.1.0_R.tar.gz", repos = NULL, type = "source")

Example R session:

library(quantbayes)
??quantbayes() # open the vignette manual
x <- as.matrix(read.table("matrix.txt"))
res <- quant_es_core(x)

Install and run the R package

From an R session, for example in RStudio:

setwd("./software_release")
install.packages("quantbayes_v0.1.0_R.tar.gz", repos = NULL, type = "source")

library(quantbayes)
set.seed(666)

??quantbayes() # open the vignette manual

# 1. LOAD DEMO DATA ####
data(core_test_data, package = "quantbayes")
print(head(core_test_data))

# Construct matrix
x <- as.matrix(core_test_data[, -1])
rownames(x) <- core_test_data[[1]]

# 2. RUN CORE FUNCTION ####
res <- quant_es_core(x)
print(res$global)
print(head(res$variants))

# if you want data frame
res_df <- as.data.frame(res$variants)
global_df <- as.data.frame(res$global)

# sort by theta_mean decreasing
res_df <- res_df[order(res_df$theta_mean, decreasing = TRUE), ]
head(res_df)
head(global_df)

# 3. CORE FUNCTION WITH CUSTOM PRIORS ####
# not recommended for our use but may be useful
res_custom <- quant_es_core(x, a = 2, b = 3, ci_level = 0.90)
print(res_custom$global)

# 4. TEST PLOTTING (default aesthetics) ####
plots <- quant_es_plots(res, x)
print(plots$p_global)
print(plots$p_overlay)
print(plots$p_matrix)
print(plots$p_p_hat)
print(plots$p_theta_ci)

# 5. FLAGSHIP PLOT ####

# Flagship overlay plot with internal colours and clean legend
swiss_red <- "#ee4035"
federal_blue <- "#2f4356"

highlight_flagship <- list(
  list(id = core_test_data[[1]][1], colour = swiss_red, size = 4),
  list(id = "6-17481874-C-G_AR", colour = federal_blue, size = 4)
)

plots_flagship <- quant_es_plots(
  res,
  x,
  highlight_points = highlight_flagship
)

flagship_plot <- plots_flagship$p_overlay +
  ggplot2::guides(
    fill = ggplot2::guide_legend(title = "highlighted variants"),
    size = "none"
  ) +
  ggplot2::labs(
    title = "GLobal posterior theta distribution",
    subtitle = "Top 10 CrI estimates with evidence available\nand two highlighted variants"
  ) +
  ggplot2::theme(
    legend.position = "right",
    legend.title = ggplot2::element_text(size = 10),
    legend.text = ggplot2::element_text(size = 9)
  )

flagship_plot

Example input matrix

The binary distribution includes example_data/test_matrix_01.txt. Each row is a variant and each column is an evidence rule. QuantBayes converts NA to zero for consistent behaviour across systems.

Support

QuantBayes runs in macOS and Linux environments, including HPC environments without administrator access. Contact the development team if you need integration support.

Licence

QuantBayes is released under the MIT Licence.

Citation

To cite quantbayes in publications, use:

The Quantitative Omics Epidemiology Group et al. (2025). “A Bayesian model for quantifying genomic variant evidence sufficiency in Mendelian disease.” medRxiv. doi: 10.64898/2025.12.02.25341503 . Preprint.

A BibTeX entry for LaTeX users is:

@Article{quantbayes2025,
  title   = {A Bayesian model for quantifying genomic variant evidence sufficiency in Mendelian disease},
  author  = {{The Quantitative Omics Epidemiology Group} and others},
  year    = {2025},
  journal = {medRxiv},
  doi     = {10.64898/2025.12.02.25341503},
  note    = {Preprint},
}