AminoScribe

AminoScribe is a Python module for generating simulated nanopore squiggle signals from amino acid sequences. It provides tools for sequence-based signal generation, time warping, noise addition, and signal processing such as filtering, normalization, and downsampling.

Features

• Generate idealized templates for amino acid sequences.
• Add time-domain warping and amplitude noise to simulate realistic signals.
• Apply low-pass Bessel filtering to reduce noise.
• Normalize signals using min-max scaling.
• Downsample signals for efficient processing.
• Fetch protein sequences using UniProt accession numbers.

Installation

Install AminoScribe using pip:

pip install aminoscribe

Usage

Generate a Squiggle Signal

You can generate a simulated squiggle signal from an amino acid sequence or a protein ID:

from aminoscribe.aminoscribe import generate_squiggle

# Generate a squiggle signal from a sequence
signal = generate_squiggle(sequence="MKTLLDLGYTMKTLLLTLVVTMKTLLDLGYTMKTLLLTLVVLLTLVVVTIVCLDLGYTLGYT", 
                           normalize=True, 
                           downsample=True, 
                           downsample_factor=5)

# Generate a squiggle signal from a protein ID
signal = generate_squiggle(protein_id="P12345", 
                           filter_noise=True, 
                           bessel_N=8, 
                           bessel_Wn=0.1)

Generate an Idealized Template

If you only need the idealized template without noise or processing:

sequence = "YYYYYSTSSDGDEEDGDDSTSYYYYYSTSSDGEDDEGDDSTSYYYYYSTSSDGEDEDGDDSTSYYYYYSTSSDGD"
template = generate_squiggle(sequence=sequence, template_only=True)

Fetch Protein Sequence

Retrieve a protein sequence using its UniProt accession number:

from aminoscribe.aminoscribe import get_protein_seq

sequence = get_protein_seq("E2RYF6")

Function Reference

generate_squiggle

Generates a simulated squiggle signal from an amino acid sequence or protein ID.

Parameters:

• sequence (str, optional): Amino acid sequence.
• protein_id (str, optional): Protein ID to fetch the sequence.
• base_template (optional): Base template signal.
• seed (optional): Random seed for reproducibility.
• template_only (bool, optional): Return idealized template only.
• cterm (str, optional): Sequence to append to the C-terminal end.
• nterm (str, optional): Sequence to prepend to the N-terminal end.
• filter_noise (bool, optional): Apply low-pass Bessel filter.
• bessel_N (int, optional): Order of the Bessel filter.
• bessel_Wn (float, optional): Normalized cutoff frequency.
• normalize (bool, optional): Apply min-max normalization.
• norm_cutoff (int, optional): Number of elements for normalization.
• downsample (bool, optional): Apply linear downsampling.
• downsample_factor (float, optional): Downsampling factor.

Returns:

• List of float values representing the processed squiggle signal.

get_protein_seq

Fetches a protein sequence using its UniProt accession number.

Parameters:

• protein_id (str): UniProt accession number.

Returns:

• Amino acid sequence as a string.

License

This project is licensed under Creative Commons Attribution-NonCommercial 4.0 International (CC BY-NC 4.0). See the LICENSE file for details.

Contributing

Source code lives at https://github.com/uwmisl/Amino-Scribe. Please submit a pull request or open an issue for any bugs or feature requests.

Contributors

Contributors

• Melissa Queen — Lead author and maintainer
• Daphne Kontogiorgos-Heintz — Computed and incorporated the amino acid value and variance numbers