MeLSI
Bioc currentMetric Learning for Statistical Inference in Microbiome Analysis
Release Lineage
Entered 3.23 · Apr 29, 2026
Current · Requires R 4.6
Description
MeLSI (Metric Learning for Statistical Inference) is a novel machine learning method for microbiome data analysis that learns optimal distance metrics to improve statistical power in detecting group differences. Unlike traditional distance metrics (Bray-Curtis, Euclidean, Jaccard), MeLSI adapts to the specific characteristics of your dataset to maximize separation between groups. The method uses an ensemble of weak learners to identify which microbial features drive group differences, providing both improved statistical power and biological interpretability through feature importance weights.
None published for Bioconductor
Test coverage
Line coverage
–
Expression
–
Tests / Examples
–
Functions
30 6 exported
Complexity
6.5 avg / 28 max
Call network
30 nodes / 27 edges
Test coverage is not measured for Bioconductor packages; nodes fall back to a neutral fill.
Call graph
Open call graph →Lowest coverage
Per-function coverage is not measured for this package yet.
Code
Structure
Lines of code
2,333
Files
22
Compiled share
9%
Has compiled src
Yes
Language breakdown
API
Exported functions
6
Internal functions
18
Recent export changes
Testing
Has tests
Yes
Test-to-code ratio
0.15
testthat edition
–
Docs
Roxygen coverage
100%
Health & Security signals
Informational signals; not verdicts.
on.exit coverage
–
Unsafe pattern score
0
Dep constraint coverage
0%
Secret pattern count
0
Bundled 3rd-party code
2 items
Portability & License
Min R version
4.5.0
System requirements
–
C++ standard
–
License
MIT + file LICENSE
License flags
SPDX valid, OSI approved
History
Versions
1
First release
2026-06-18
Latest release
2026-06-18
Avg cadence
–
Cold removal rate
–
Dep drift
0
Per-file churn detail lives in the source pipeline: https://github.com/r-observatory/bioc-code-metrics.
Documentation
- Help pages
- 7
- Examples that run
- 100%7 run in a plain check
- Exported functions with examples
- 100%
- Documented parameters
- 100%
- Return-value docs
- 100%
- Help pages citing references
- 0%
- MeLSI Tutorial HTML
- Release backport of the cumulative performance work developed in the 1.1.x devel series, brought to the Bioconductor release branch. Two fused C++ (Rcpp) kernels replace the R hot paths: the PERMANOVA pseudo-F statistic and the weak-learner metric optimizer (
src/melsi_fstat.cpp). Together with the R-level optimizations (vectorized prefiltering, direct F-statistic in place ofvegan::adonis2, diagonal-metric column scaling, diagonal-vector ensemble storage), this speeds up analysis substantially (up to ~100x on high-dimensional inputs) with no change to the statistical method. - Reproducibility: results are bit-faithful to 1.0.0. The C++ optimizer draws its class pairs with R's own index sampler (
R_unif_index), consuming the global RNG stream identically to the previoussample()-based loop; F-statistics reproduce to floating-point precision with no p-value changes. Validated against the published simulation tables (Type I error, scalability, parameter sensitivity, feature correlation).
From NEWS.md in the release
Repository practices
2 development-tooling and community-health practices detected across 1 family in the upstream repository
Checks run against github.com/nathanbresette/melsi on 2026-10-18.
The R CMD check workflow runs on Linux
Topics
People
- Nathan Bresette author maintainer
- Aaron C. Ericsson author
- Ai-Ling Lin author funder
- Carter Woods author
Cite
Cite this package
Run in R for the authors' entry:
citation("MeLSI")The package provides a citation file, so citation() returns the authors' own entry.
It points to 10.5281/zenodo.17714848, 10.64898/2025.12.04.692328.
Cite the R Observatory
For a number measured here: a download total, a coverage figure, an archival date.
From data release v2026-10-21, which the citation names so these numbers can be found later. More on citing and the projects behind them.