GENOMICS-CDDBRG/SeqPure-BLAST

1---2title: Bulamarwa Metagenomics Tool 23emoji: 📊4colorFrom: blue5colorTo: green6sdk: docker7app_port: 78608pinned: false9license: mit10---11 12# Bulamarwa Metagenomics Tool 2: Statistics and Reporting13 14Second tool in a two tool metagenomics workflow. Takes a species abundance table15and sample metadata, typically `species_relative_abundance.tsv` (or16`species_estimated_reads.tsv` for count based Chao1) and17`samplesheet_resolved.csv` produced by **Gulumba Metagenomics Tool 1**.18Abundance tables from other pipelines, such as MetaPhlAn or HUMAnN, are also19accepted in either counts or relative abundance form.20 21## What it does22 23- Alpha diversity: observed richness, Shannon, Simpson, and Chao1 when a24  validated integer count matrix is supplied.25- Beta diversity: Bray-Curtis, Jaccard, or Aitchison distance with PCoA.26- PERMANOVA and PERMDISP, both computed by direct label permutation rather27  than a parametric approximation.28- Exploratory CLR differential abundance with Benjamini-Hochberg FDR29  correction, explicitly labelled as exploratory rather than a30  covariate-adjusted ANCOM-BC2 or MaAsLin2 substitute.31- Optional cross-validated Random Forest classification, with out-of-fold32  performance metrics, held-out permutation importance, and post-hoc SHAP33  explanation.34- A `publication_package.zip` with a `tables/` folder, a `figures/` folder at35  300 dpi with editable SVGs, a `methods.txt` you can adapt for a manuscript,36  and a `README.txt` describing every file.37 38## Deploying this Space39 401. Create a new Space on Hugging Face, select the **Docker** SDK.412. Upload `app.py`, `requirements.txt`, and `Dockerfile` from this package42   (this `README.md` can go too, since its front matter configures the Space).433. Hugging Face builds the image and starts the container automatically.44   No external database is required for this tool.45 46## Local test before deploying47 48```bash49pip install -r requirements.txt50python app.py51```52 53The app listens on port 7860 by default, matching `app_port` above.54 55## Notes56 57- All computation happens inside the container. No abundance data is sent58  anywhere outside the Space.59- Machine learning results are only meaningful with enough samples per group;60  the tool will skip that step rather than report an unreliable score if the61  dataset is too small.62