A Systems-Level DNA Storage Archiver written in Rust.

Streaming I/O • AES-256-GCM • Reed-Solomon (N+K) • Viterbi Correction

Helix is a high-performance compiler designed to bridge the gap between binary data and biological storage. It transforms digital files into biostable DNA oligonucleotides formatted for synthesis and deep-time archival.

Unlike simple transcoders, Helix implements a full "Systems Storage" stack, handling encryption, error correction (Erasure + Viterbi), biological safety checks, and random access retrieval.

Important

Experimental Research Prototype This project is a rigorous software implementation of DNA storage principles. While the encoding algorithms are designed to be biologically sound (GC-balanced, homopolymer-free, collision-resistant), this specific implementation has not yet been validated via wet-lab synthesis and sequencing.

Use this tool for research, simulation, and algorithmic verification. Do not use for critical long-term archival without physical validation of the primer sets and payload stability.

---

🚀 Key Capabilities

⚡ Performance & Scale

• Smart Streaming Architecture: * Constant Memory Footprint: Processes files in 4MB streaming chunks. This allows archiving multi-terabyte datasets with a minimal RAM footprint (~80MB peak), preventing OOM crashes even on constrained legacy hardware.
  • Memory-Aware Backpressure: The batch iterator monitors byte usage, not just line counts, ensuring "DNA Soup" files (massive single lines or many small lines) never exhaust physical RAM.
• Massively Parallel: Utilizes Rayon to parallelize CRC hashing, Reed-Solomon encoding, DNA translation, search filtering, and decay simulation across all available CPU cores (-j flag).
• Zstd Compression: Applies Zstandard (Level 3) compression before encoding to maximize the Bits-per-Molecule density.

🧬 Biological Integrity

• Homopolymer Prevention: Uses a Rotating Base-3 Trellis state machine. This ensures that no base is ever repeated (e.g., AAAA or GGGG is mathematically impossible), significantly reducing sequencing errors.
• Auto-Correction for Stability: * Salt & Retry Mechanism: If a block produces unstable DNA (bad GC content or $T_m$), the compiler automatically rotates the block's cryptographic salt and re-encodes. This changes the bitstream—and thus the DNA sequence—transparently until biological constraints are met.
  • Synthesis Safety Guard: Analyzes every strand for GC-Content (40-60% window) and Melting Temperature ($T_m$).
• Fuzzy Primer Matching: The decoder employs Hamming distance checks (tolerance of 3 mismatches) to identify primers even when mutated. This prevents valid data from being discarded due to "Zip Code" rot.
• Primer Collision Avoidance: Scans payloads for accidental primer sequences and utilizes trellis chaining (FP -> Address -> Payload -> RP) to ensure seamless transitions.

🛡️ Security & Resilience

• Cryptographic Access: * Argon2id for Master Key derivation (memory-hard).
  • HKDF + AES-GCM for per-block session keys. A unique nonce and salt for every block means identical files produce completely different DNA streams.
• Multi-Layer Error Correction:
  • Reed-Solomon (Erasure Coding): Configurable redundancy (Default: 10 Data + 5 Parity) recovers files even if 33% of strands are completely lost.
  • Viterbi Decoder (Mutation Correction): Treats DNA as a "Noisy Channel." If a strand fails integrity checks, the Viterbi engine finds the optimal path through the trellis to "heal" substitution errors, recovering data from strands with ~1.0% mutation rates.
• Chemical Corruption Detection: A CRC32 checksum is prepended to every shard to validate the final output of the Viterbi decode.

🔍 Molecular Random Access

• In-Silico PCR (Streaming Search): Supports memory-safe "Soft-Search" by filtering gigabytes of mixed DNA data ("The Soup") for specific primer tags using a parallelized, streaming map-reduce approach.
• Configurable Primers: Users can define custom Forward/Reverse primers to physically address specific files within a biological pool.

---

🏗 System Architecture

The Helix Pipeline operates on 4MB independent blocks, transforming binary data through 5 distinct layers:

1. L1 - Stream & Compress: The file is read in buffered 4MB chunks and compressed via Zstd.
2. L2 - Encryption: The compressed chunk is encrypted (AES-256-GCM) using a unique nonce and salt per block. Note: If stability checks fail, this step is re-run with a new salt.
3. L3 - Redundancy: The blob is split into $N$ data shards. $K$ parity shards are generated using Galois Field arithmetic (Reed-Solomon).
4. L4 - Transcoding: * Each shard is prepended with a CRC32 checksum.
   • Binary data is mapped to DNA bases using the constrained trellis.
   • Primers and Index Addresses are attached: [FwdPrimer] [Address] [Payload] [RevPrimer].
5. L5 - Analysis: The resulting Oligo is checked for biological stability metrics (GC% and $T_m$).

---

📦 Installation

# Clone the repository
git clone https://github.com/SSL-ACTX/helix.git
cd helix

# Build optimized binary
cargo build --release

# Or install directly
cargo install --git https://github.com/SSL-ACTX/helix.git

---

💻 Usage Guide

1. Compile (Archive)

Encrypts, compresses, and encodes a file into a DNA stream.

# Standard encoding (Auto-threading)
./target/release/helix compile database.dump --output archive.fasta

# High-Security Mode (Custom Password & High Redundancy)
./target/release/helix compile secrets.pdf \
    --password "hunter2" \
    --data 20 --parity 10

# Custom Primers (for physical PCR addressing)
./target/release/helix compile project.zip \
    --primer-fwd "GCTAGCTAGCTAGCTAGCTA" \
    --primer-rev "CGATCGATCGATCGATCGAT"

2. Search (Molecular Filtering)

Extracts specific strands from a massive DNA dataset based on tags or primers. Now safe for files larger than RAM.

# Search by Tag
./target/release/helix search soup.fasta "project_alpha" --output found.fasta

# Search by Custom Primer
./target/release/helix search soup.fasta \
    --primer-fwd "GCTAGCTAGCTAGCTAGCTA" \
    --primer-rev "CGATCGATCGATCGATCGAT" \
    --output found.fasta

3. Restore (Decode)

Recovers the binary file from a DNA stream. Supports out-of-order recovery and streaming writes.

./target/release/helix restore archive.fasta recovered.file \
    --password "hunter2" \
    --data 20 --parity 10

4. Simulate Decay (Chaos Monkey)

Simulates "Deep Time" storage by randomly deleting strands (dropout) and introducing bit-rot (mutation) to test robustness.

# Simulate 10,000 years of decay (30% dropout + 0.5% mutation rate)
./target/release/helix simulate archive.fasta \
    --dropout 30 \
    --mutation 0.005 \
    --output decayed.fasta

---

🧪 Verification

Helix includes a rigorous Python validation suite (full_test.py) that tests the entire stack against edge cases:

• Concurrency Interop: Verifies thread safety between sequential and parallel modes.
• Cryptographic Denial: Ensures wrong passwords yield fatal errors.
• Catastrophic Data Loss: Tests recovery limits (> Parity limit).
• Bit-Rot/Mutation: Verifies CRC32 detection of mutated bases using the internal mutation simulator.
• Viterbi Repair: Validates the dynamic programming engine against heavy mutation scenarios (1.0% error rate).
• Stability Enforcement: Stresses the "Salt & Retry" engine with pathological binary inputs.
• Primer Safety: Fuzzing tests to ensure no accidental primer collisions occur in the payload.
• Streaming Stress: Validates multi-block processing with files > RAM.

To run the full suite:

python tests/full_test.py

---

**Built with 🦀 and ☕ by Seuriin**