sparcc

name	sparcc
description	SparCC — Sparse Correlations for Compositional data; infer taxon–taxon correlations from relative abundances without naive Pearson on compositions. Use for correlation networks on microbiome tables. Original: https://bitbucket.org/yonatanf/sparcc; convenient R implementation inside SpiecEasi (sparcc()). Route via research-analysis; conditional networks → spieceasi.
license	MIT
category	evaluation
tags	["SparCC","network","correlation","compositionality","microbiome"]
upstream	https://bitbucket.org/yonatanf/sparcc
stage	report

SparCC

Upstream (original): https://bitbucket.org/yonatanf/sparcc · R helper often via SpiecEasi sparcc()

Citation

Friedman, J. & Alm, E. J. Inferring correlation networks from genomic survey data. PLoS Comput. Biol. 8, e1002687 (2012). https://doi.org/10.1371/journal.pcbi.1002687

See also docs/references.md.

Analytical thinking

Relative abundances are compositional — Pearson/Spearman on proportions yields spurious correlations. SparCC estimates correlations under a sparse assumption. Correlation ≠ direct ecological interaction; for conditional independence graphs prefer spieceasi. Filter rare taxa; report bootstrap / thresholding if used.

How to run

# via SpiecEasi (common)
# remotes::install_github("zdk123/SpiecEasi")
library(SpiecEasi)
# sparcc.amgut <- sparcc(counts_matrix)
# threshold abs(Cor) for edges — pin cutoff in Methods

Or follow the original SparCC Python/bitbucket pipeline for the installed tag.

Decision tree

Microbial association network?
├─ Compositional correlations → sparcc
├─ Sparse conditional networks → spieceasi
├─ DA between groups → ancombc / maaslin2
└─ Plot networks in R suite → microeco (± igraph)

Related skills

spieceasi · research-analysis · microeco · phyloseq · maaslin2 · tool-selection