Introduction to MAGICR

Overview

MAGICR is a native R implementation of the MAGIC (Markov Affinity-based Graph Imputation of Cells) algorithm for denoising and imputation of single-cell RNA sequencing (scRNA-seq) data.

Single-cell RNA sequencing has revolutionized our understanding of cellular heterogeneity, but technical limitations result in sparse count matrices with prevalent dropout events—where expressed genes appear as zeros due to sampling inefficiency. MAGIC addresses this challenge by leveraging the underlying manifold structure of the data through diffusion geometry.

Installation

# From R-universe (recommended)
install.packages("MAGICR", repos = "https://zaoqu-liu.r-universe.dev")

# From GitHub
remotes::install_github("Zaoqu-Liu/MAGICR")

Quick Start

library(MAGICR)

# Load example data
data(magic_testdata)

# Check data dimensions
dim(magic_testdata)
#> [1] 500 100

# Preview the data
magic_testdata[1:5, 1:5]
#>        Gene_1 Gene_2 Gene_3 Gene_4 Gene_5
#> Cell_1      8      3      7      2     11
#> Cell_2      9      3      2      9      4
#> Cell_3      4      5      7      1      6
#> Cell_4      7      7      5      8      8
#> Cell_5      6      2      5      4      7

Running MAGIC

# Run MAGIC with default parameters
result <- magic(magic_testdata, t = 3)

# View result summary
print(result)
#> MAGIC Result
#> ============
#>   Cells: 500
#>   Genes: 100
#>   Parameters:
#>     knn: 5
#>     decay: 1
#>     t: 3
#>     solver: exact
#>     knn_dist: euclidean
#>     npca: 100

Accessing Results

# Get imputed matrix
imputed_data <- as.matrix(result)

# Compare original vs imputed
cat("Original data range:", range(magic_testdata), "\n")
#> Original data range: 0 17
cat("Imputed data range:", range(imputed_data), "\n")
#> Imputed data range: 4.375909 5.699832

Visualizing Results

Before vs After Imputation

par(mfrow = c(1, 2))

# Original data distribution
hist(as.vector(as.matrix(magic_testdata)), 
     breaks = 50, 
     main = "Original Data Distribution",
     xlab = "Expression", 
     col = "#3498db", 
     border = "white")

# Imputed data distribution
hist(as.vector(imputed_data), 
     breaks = 50, 
     main = "Imputed Data Distribution",
     xlab = "Expression", 
     col = "#e74c3c", 
     border = "white")

Gene-Gene Relationships

One of MAGIC’s key benefits is recovering gene-gene relationships that are obscured by dropout noise.

# Select two genes
gene1 <- colnames(magic_testdata)[1]
gene2 <- colnames(magic_testdata)[2]

par(mfrow = c(1, 2))

# Original
plot(magic_testdata[, gene1], magic_testdata[, gene2],
     pch = 16, col = adjustcolor("#3498db", 0.5),
     xlab = gene1, ylab = gene2,
     main = "Original")

# Imputed
plot(imputed_data[, gene1], imputed_data[, gene2],
     pch = 16, col = adjustcolor("#e74c3c", 0.5),
     xlab = gene1, ylab = gene2,
     main = "After MAGIC")

Parameter Tuning

Diffusion Time (t)

The diffusion time t controls the degree of smoothing:

Small t (1-3): Less smoothing, preserves more local structure
Large t (>5): More smoothing, may over-smooth rare populations
“auto”: Automatically selects optimal t using Procrustes analysis

# Automatic t selection
result_auto <- magic(magic_testdata, t = "auto", t_max = 10)
cat("Automatically selected t:", result_auto$params$t, "\n")
#> Automatically selected t: 10

Key Parameters

Parameter	Default	Description
`knn`	5	Neighbors for bandwidth estimation
`knn_max`	15	Maximum neighbors for graph
`decay`	1	Kernel sharpness (α parameter)
`t`	3	Diffusion time
`npca`	100	PCA components for distances

Session Info

sessionInfo()
#> R version 4.4.0 (2024-04-24)
#> Platform: aarch64-apple-darwin20
#> Running under: macOS 15.6.1
#> 
#> Matrix products: default
#> BLAS:   /Library/Frameworks/R.framework/Versions/4.4-arm64/Resources/lib/libRblas.0.dylib 
#> LAPACK: /Library/Frameworks/R.framework/Versions/4.4-arm64/Resources/lib/libRlapack.dylib;  LAPACK version 3.12.0
#> 
#> locale:
#> [1] C
#> 
#> time zone: Asia/Shanghai
#> tzcode source: internal
#> 
#> attached base packages:
#> [1] stats     graphics  grDevices utils     datasets  methods   base     
#> 
#> other attached packages:
#> [1] MAGICR_1.0.0
#> 
#> loaded via a namespace (and not attached):
#>  [1] cli_3.6.5         knitr_1.51        rlang_1.1.7       xfun_0.56        
#>  [5] otel_0.2.0        textshaping_1.0.4 jsonlite_2.0.0    listenv_0.10.0   
#>  [9] htmltools_0.5.9   ragg_1.5.0        sass_0.4.10       rmarkdown_2.30   
#> [13] grid_4.4.0        evaluate_1.0.5    jquerylib_0.1.4   fastmap_1.2.0    
#> [17] yaml_2.3.12       lifecycle_1.0.5   compiler_4.4.0    codetools_0.2-20 
#> [21] irlba_2.3.5.1     fs_1.6.6          Rcpp_1.1.1        htmlwidgets_1.6.4
#> [25] future_1.69.0     lattice_0.22-7    systemfonts_1.3.1 digest_0.6.39    
#> [29] R6_2.6.1          RANN_2.6.2        parallelly_1.46.1 parallel_4.4.0   
#> [33] bslib_0.9.0       Matrix_1.7-4      tools_4.4.0       globals_0.18.0   
#> [37] pkgdown_2.1.3     cachem_1.1.0      desc_1.4.3

Zaoqu Liu

2026-01-26