ConsensusCluster

An R package for performing k-means consensus clustering with bootstrap resampling.

Installation

You can install the package directly from this repository using devtools:

# Install devtools if you haven't already
install.packages("devtools")

# Install ConsensusCluster
devtools::install_github("skelly001/ConsensusCluster")

Or install locally:

devtools::install("path/to/ConsensusCluster")

Quick Start

library(ConsensusCluster)

# Create sample data
set.seed(123)
data_matrix <- matrix(rnorm(1000), nrow = 100, ncol = 10)
rownames(data_matrix) <- paste0("Feature", 1:100)

# Run consensus clustering
clusters <- kmeansCC(
  data = data_matrix,
  k = 4,
  reps = 100,
  ncores = 4,
  seed = 0
)

# View cluster assignments
table(clusters)

Function: kmeansCC

Performs consensus clustering using k-means algorithm with bootstrap resampling.

Parameters

• data: A numeric matrix or data frame with features in rows and samples in columns
• k: Integer. The number of clusters to identify
• reps: Integer. The number of bootstrap iterations to perform (default: 100)
• ncores: Integer. The number of CPU cores to use for parallel processing (default: 1)
• seed: Integer. Random seed for reproducibility (default: 0)

Returns

A named integer vector of cluster assignments for each feature.

Algorithm

1. Generate all unique pairs of features (including self-pairs)
2. For each bootstrap iteration:
   • Resample samples with replacement
   • Perform k-means clustering
   • Record which features cluster together
3. Calculate consensus proportions: how often each pair of features co-clustered
4. Build a symmetric co-clustering matrix
5. Perform hierarchical clustering on 1 - consensus matrix
6. Cut tree to obtain k final clusters

License

MIT License