mirnaprotpred 0.1.7

A pip installable CLI for miRNAProtPred tool

miRNAProtPred

A powerful, high-performance bioinformatics framework for discovering, evaluating, and verifying microRNA (miRNA) interactions across DNA, RNA, and protein target sequences.

The mirnaprotpred package provides two core modules:

1. SeqFinder: A discovery engine to find all potential miRNA interactions across a genome or target sequence.
2. Validator: A targeted verification engine to test specific, user-provided miRNAs against a target sequence.

Both modules are powered by a shared, rigorous biological engine that evaluates exact seed matching, wobble pairing, AU-rich context, and RNAduplex thermodynamic stability.

Features

• Multi-Sequence Support: Natively processes DNA, RNA, and Protein sequences. Protein sequences are automatically reversed-mapped to nucleotides via an integrated BLAST search. Note: Protein mode provides exploratory nucleotide-region inference and should not be interpreted as direct protein-target prediction.
• Strict & Relaxed Modes:
  • strict: Identifies only canonical, exact seed interactions.
  • relaxed: Discovers flexible interactions including wobble pairings, with biological confidence re-weighting.
• Thermodynamic Scoring: Leverages ViennaRNA (RNAduplex) to calculate Minimum Free Energy (MFE) structural stability.
• Multi-Factor Confidence Classification: Interactions are graded (Very High, High, Medium, Low) based on MFE, motif identity, AU-context accessibility, and cluster density.
• Targeted Validation: Provide specific miRNA IDs (inline or via text/FASTA files) to rapidly check for interactions without scanning the entire database.
• Flexible Output Options: Terminal-friendly concise summaries, full detailed interaction data, high-confidence filtering, and direct CSV export.

Installation

Prerequisites

• Python 3.9 or higher
• ViennaRNA (Must be installed via conda or built from source)
• BLAST+ (Optional, required only for local protein BLAST if not using NCBIWWW)

Install via Conda (Recommended)

To ensure ViennaRNA installs correctly, using Conda is highly recommended:

conda create -n mirnaprotpred_env python=3.9
conda activate mirnaprotpred_env
conda install -c bioconda viennarna

Install from Source

Important: You must still install ViennaRNA via Conda before installing from source, as it cannot be natively compiled via pip.

conda install -c bioconda viennarna
git clone https://github.com/somenath-combio/mirnaprotpred.git
cd mirnaprotpred
pip install -e .

Usage: SeqFinder

Use SeqFinder to scan an entire sequence and discover all potential miRNA interactions.

SeqFinder <target_sequence_or_fasta> [options]

Examples

# Scan a FASTA genome file
SeqFinder examples/sars.fasta

# Relaxed mode to include wobble pairings
SeqFinder examples/sars.fasta --mode relaxed

# View detailed raw output for all interactions
SeqFinder examples/sars.fasta --output raw

# View only High and Very High confidence interactions
SeqFinder examples/sars.fasta --output highconf

# Save results automatically to CSV
SeqFinder examples/sars.fasta --output raw --out results.csv

# Direct sequence input
SeqFinder "AUGCAUGCAUGCAUGC"

Usage: Validator

Use the validator to check if specific miRNAs interact with your target.

validator <miRNA_IDs> <target_sequence_or_fasta> [options]

The validator provides a simple YES/NO summary by default, but can output the full SeqFinder interaction data using --details.

Examples

# Inline comma-separated miRNAs
validator hsa-miR-21-5p,hsa-miR-122-5p examples/sars.fasta

# Read miRNAs from a text file (one per line)
validator examples/mirnas.txt examples/sars.fasta

# Extract miRNA IDs from a FASTA file and validate them
validator known_mirnas.fasta examples/sars.fasta

# Show full detailed interaction data for the matched miRNAs
validator examples/mirnas.txt examples/sars.fasta --details

# Save results automatically to CSV
validator examples/mirnas.txt examples/sars.fasta --out validation.csv

How It Works

1. Input Processing: The engine detects whether the input is DNA, RNA, or Protein. If Protein, a tblastn search is performed against the NCBI database to retrieve the corresponding nucleotide region.
2. Motif Discovery: The Boyer-Moore string matching algorithm rapidly identifies potential binding sites using canonical or relaxed seed variants.
3. Contextual Evaluation: The 20nt flanking regions are evaluated for AU content, which correlates with structural accessibility.
4. Structural Analysis: ViennaRNA calculates the optimal secondary structure and minimum free energy (MFE) between the miRNA and the target site.
5. Clustering & Ranking: Competing seeds at the same locus are clustered, scored using a composite biological metric, and the optimal interaction is selected.
6. Confidence Classification: Results are categorized based on stringent biological thresholds (e.g., Very High confidence requires strict identity, strong MFE, and high AU context).

Output Columns (Detailed Mode)

• miRNA_ID / description: The matched human miRNA.
• CTS: Complementary Target Sequence (the matching region on your input).
• CTS_start: 1-based start coordinate of the interaction.
• RNAduplex_MFE: Thermodynamic stability (Minimum Free Energy, more negative is stronger).
• AU_Context: Accessibility score based on AU richness in flanking regions.
• motif_identity / match_type: Structural identity of the seed interaction (Exact vs Wobble).
• Confidence_Level: Biological confidence (Very High, High, Medium, Low).

Authors

• Somenath Dutta ([e-mail])
• Sudipta Sardar ([e-mail])

Citation

If you use this tool in your research, please cite: (Citation details pending publication)

License

MIT License