sceptr-profiling 1.0.0

Statistical Characterisation of Expression Profiles in Transcriptomes

SCEPTR

Statistical Characterisation of Expression Profiles in Transcriptomes

A statistical framework for continuous enrichment profiling of ranked transcriptomes. SCEPTR computes enrichment functions E_C(k) at every gene rank, permutation-based significance testing, and D_KL functional specialisation gradients - all from a single sample with no replicates required.

Installation

pip install sceptr

For GO hierarchy support:

pip install sceptr[go]

Quick Start

# List available category sets
sceptr categories --list

# Run enrichment profiling
sceptr profile --expression my_data.tsv --category-set bacteria -o results/

Input formats

SCEPTR accepts several input formats:

1. Annotated expression table (recommended) - a TSV with gene IDs, expression values, and protein descriptions:

sequence_id    TPM    protein_name    GO_Biological_Process
gene001        2500   Ribosomal protein L3    translation [GO:0006412]
gene002        1800   ATP synthase subunit    ATP synthesis [GO:0015986]

2. Expression + external category mapping - when you have your own category assignments:

sceptr profile --expression expr.tsv --categories mapping.tsv -o results/

Where mapping.tsv is:

gene_id    category
gene001    Translation & Ribosome
gene001    Protein Folding
gene002    Central Metabolism

3. Pre-categorised expression table - a TSV with a categories column:

sequence_id    TPM    categories
gene001        2500   Translation & Ribosome;Protein Folding
gene002        1800   Central Metabolism

Custom categories

You can define your own category sets as JSON:

sceptr profile --expression data.tsv --custom-categories my_categories.json -o results/

What SCEPTR computes

• Continuous enrichment profiles E_C(k) at every integer gene rank
• Discrete tier enrichment with Fisher's exact test and FDR correction
• Permutation-based global profile test (supremum and integral statistics)
• D_KL functional specialisation gradient quantifying transcriptome organisation
• Profile shape classification (apex-concentrated, distributed, flat)
• Interactive HTML report with all results in a single self-contained file

Python API

from sceptr.profile import run

results = run(
    expression_file="data.tsv",
    category_set="bacteria",
    output_dir="results/",
    permutations=1000,
)

# Access results programmatically
tier_results = results["tier_results"]
continuous = results["continuous"]
k_values = continuous["k_values"]
enrichment_matrix = continuous["enrichment_matrix"]

Citation

McCabe, J.S. and Janouskovek, J. (2026). SCEPTR: continuous enrichment profiling reveals functional architecture across the expression gradient.

Full framework

For end-to-end analysis from raw reads (QC, quantification, annotation, and profiling), see the full SCEPTR framework which uses Nextflow and Docker.