---
title: FtsI Classifier
emoji: 📈
colorFrom: gray
colorTo: red
sdk: gradio
sdk_version: 6.1.0
app_file: app.py
pinned: false
license: mit
short_description: Finding unique k-mer signatures among groups of sequences
---

# Genomic Classification of Beta-Lactam Resistance

This repository provides a Python implementation of a genomic classification framework based on k-mer analysis of gene sequences (ftsI / penA) to predict beta-lactam susceptibility categories.

## Description

The method identifies group-specific k-mers from allele sequences and uses them to classify isolates into resistance categories. The approach is designed to capture genomic patterns beyond predefined mutation rules.

This framework was developed and applied to:
- *Haemophilus influenzae* (ftsI gene)
- *Neisseria meningitidis* (penA gene)

## Files

- `app.py` — main application (Gradio interface)
- `kmer_unique.py` — core k-mer extraction and filtering functions
- `requirements.txt` — Python dependencies

## Requirements

Python ≥ 3.9

Install dependencies:

```bash
pip install -r requirements.txt
## Usage

### Hugging Face interface

Run the app directly via the Hugging Face Space interface.

## Input

- FASTA file containing gene sequences (ftsI or penA)
- Sequences grouped by phenotype or category

## Output

- Predicted resistance category per sequence
- Identification of group-specific k-mer signatures

## Method overview

1. Extraction of k-mers from sequences  
2. Identification of unique k-mers per category  
3. Filtering based on frequency thresholds  
4. Classification of sequences based on k-mer matches  

## Citation

If you use this code, please cite:

[to be added manuscript reference]

## DOI

DOI: (10.5281/zenodo.19665426)