spieceasi

name	spieceasi
description	SPIEC-EASI — Sparse Inverse Covariance Estimation for Ecological Association Inference; compositionally robust microbial network inference (MB / glasso) with optional SparCC. Use for conditional association networks from count / relative abundance tables. Upstream: https://github.com/zdk123/SpiecEasi. Route via research-analysis; pure SparCC correlations → sparcc.
license	MIT
category	evaluation
tags	["SPIEC-EASI","SpiecEasi","network","glasso","microbiome","compositionality"]
upstream	https://github.com/zdk123/SpiecEasi
stage	report

SPIEC-EASI (SpiecEasi)

Upstream: zdk123/SpiecEasi

Citation

Kurtz, Z. D. et al. Sparse and compositionally robust inference of microbial ecological networks. PLoS Comput. Biol. 11, e1004226 (2015). https://doi.org/10.1371/journal.pcbi.1004226

See also docs/references.md.

Analytical thinking

SPIEC-EASI infers sparse graphical models (neighborhood selection or glasso) after compositional transforms — edges closer to conditional dependence than raw correlations. Includes a sparcc() implementation for comparison. Needs enough samples relative to taxa; filter rare features; pin method (mb/glasso), λ path, and StARS/pulsar settings.

Networks are hypotheses — not proven interactions.

How to run

# remotes::install_github("zdk123/SpiecEasi")
library(SpiecEasi)
# se <- spiec.easi(counts, method='mb', lambda.min.ratio=1e-2, nlambda=20,
#                  pulsar.params=list(rep.num=50))
# extract adjacency / plot with igraph — follow README

Decision tree

Network inference need?
├─ Conditional sparse network → spieceasi
├─ SparCC correlations only → sparcc
├─ Group DA → ancombc / maaslin2
└─ Exploratory plots → microeco / phyloseq

Related skills

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