FastaPasta

FastaPasta is a beautiful and fast Tauri + Next.js desktop app for viewing, analyzing, and understanding FASTA-formatted genetic sequences.

It features:

• Sequence rendering with colored base and amino acid maps,
• Intelligent codon translation and polypeptide extraction,
• AI-based functional predictions powered by Ollama + InterPro lookups,
• A slick, dark-mode-ready interface

---

Features

• View and page through multiple FASTA entries
• Interactive color-coded sequences and amino acids
• Predict polypeptide function using local LLM (via Ollama)
• Base distribution pie charts
• Codon translation & stop-codon-based polypeptide chain extraction
• Built with Next.js, Tailwind, TypeScript, Rust, and Tauri

---

Tech Stack

Layer	Stack
Frontend	Next.js, TypeScript, Tailwind CSS
Backend	Rust (Tauri)
AI Assistant	Ollama (Local LLM server) + InterPro API
Desktop App	Tauri

---

Getting Started

Prerequisites

• Node.js
• Rust + Cargo
• Tauri CLI: cargo install tauri-cli
• Ollama (for LLM predictions): ollama.com

---

Run Locally

1. Install frontend dependencies:

   cd frontend
   npm install
   

2. Start development server:

   npm run dev
   

3. Run the Tauri app:

   cd ..
   cargo tauri dev
   

---

How It Works

• FASTA files are parsed in Rust & visualized in React.
• Sequences are color-coded based on nucleotide/amino acid types.
• Codons are translated, polypeptides are extracted.
• Users can click a chain, and the app sends the sequence to InterPro and an LLM (via Ollama) for bioinformatic insight.
• Most processing is local — only a single call of a selected polypeptide sequence is queried to EBI's InterPro API.

---

File Structure (Simplified)

├── frontend/ # Next.js frontend
├── src-tauri/ # Rust + Tauri backend
├── genomes/ # Sample FASTA files from
├── docs/ # Screenshots
├── combine_files.sh # Optional helper script

---

Notes

• Ollama must be running for AI predictions to work: ollama serve (FastaPasta will spin up the server process if needed)
• Model list is retrieved via ollama list — make sure you have at least one model installed (llama2, deepseek, gemma, etc.)
• InterPro results are fetched live from EBI's API.

---

Screenshots

Base Overview	Amino Acids View	Prediction UI

---

Resources

fasta.bioch.virginia.edu/fasta/fasta_list.html

ebi.ac.uk/jdispatcher

bioinformatics.nl/tools/crab_fasta.html

genome.ucsc.edu/cgi-bin/hgTables (Source for the example FASTA files in src-tauri/genomes.)

uniprot.org/uniprotkb

ebi.ac.uk/proteins/api/doc/index#!/proteins/search

huggingface.co/facebook/esm2_t33_650M_UR50D

wormbase.org/#012-34-5

---

Contributing

Pull requests welcome! Here's a rough roadmap of desired features:

• Add FASTA file drag-and-drop
• Add zoom/scale options for huge sequences
• Expand AI mode to allow user prompt editing
• Save/export analyzed data (CSV, JSON)