Unified platform for genetic engineering: CRISPR design, protein engineering, synthetic biology, gene therapy, computational genomics, metabolic engineering, single-cell genomics, epigenomics, AI/ML, and biosecurity.
graph TB
subgraph Input["Input Layer"]
DNA[DNA/RNA Sequences]
PROT[Protein Sequences]
OMICS[Omics Data]
end
subgraph Core["Core Processing Layer"]
CRISPR[CRISPR Design<br/>Cas9/Cas12a/Cas13<br/>Base/Prime Editing]
PROTEIN[Protein Engineering<br/>Structure/Stability]
SYN[Synthetic Biology<br/>Circuits/Pathways]
GEN[Genomics<br/>Variants/Assembly]
GT[Gene Therapy<br/>Vectors/Delivery]
MET[Metabolic Eng<br/>FBA/Strain Opt]
AIML[AI/ML<br/>Protein LMs<br/>Variant Effect]
end
subgraph Integration["Integration Layer"]
KG[Knowledge Graph<br/>SPARQL/GraphML]
EB[EventBus<br/>Pub/Sub]
DT[Digital Twin<br/>Simulation]
PIPE[Pipeline<br/>Compose/Async]
end
subgraph Security["Biosecurity Layer"]
SCREEN[Sequence Screening<br/>Threat Patterns]
BSL[BSL Classification<br/>1-4 Clearance]
AUDIT[Audit Trail<br/>Logging]
WM[Watermarking<br/>Provenance]
end
subgraph Output["Output Layer"]
DESIGN[Designs]
PRED[Predictions]
REPORT[Reports]
ALERT[Alerts]
end
DNA --> CRISPR
DNA --> GEN
PROT --> PROTEIN
OMICS --> AIML
DNA --> SYN
DNA --> GT
OMICS --> MET
CRISPR --> KG
PROTEIN --> KG
GEN --> KG
SYN --> EB
GT --> EB
MET --> DT
KG --> SCREEN
EB --> BSL
DT --> AUDIT
SCREEN --> WM
SCREEN --> ALERT
BSL --> ALERT
KG --> REPORT
DT --> REPORT
CRISPR --> DESIGN
PROTEIN --> PRED
GEN --> PRED
sequenceDiagram
participant U as User
participant P as Pipeline
participant C as CRISPR
participant B as Biosecurity
participant KG as KnowledgeGraph
participant DT as DigitalTwin
U->>P: Run workflow
P->>C: Design gRNA
C-->>P: gRNA + efficiency
P->>B: Screen sequence
B-->>P: Risk assessment
alt Passes screening
P->>KG: Store design
P->>DT: Simulate effect
DT-->>P: Predicted outcome
P-->>U: Results
else Fails screening
B-->>U: Alert + reason
end
graph LR
subgraph NP-Hard["NP-Hard Problems"]
PF[Protein Folding<br/>Unger & Moult 1993]
PD[Protein Design<br/>Pierce & Winfree 2002]
GA[Genome Assembly<br/>Medvedev 2007]
MSA[Multiple Sequence Alignment<br/>Wang & Jiang 1994]
HA[Haplotype Assembly<br/>MEC Formulation]
MPO[Metabolic Pathway Opt<br/>MILP]
GCD[Gene Circuit Design<br/>Combinatorial]
PTD[Phylogenetic Tree<br/>Reconstruction]
RNA[RNA Inverse Folding]
end
subgraph Solvers["Our Solvers"]
PF --> ESM[ESM/AlphaFold<br/>Embedding-based]
PD --> KMER[k-mer LM<br/>Scoring]
GA --> GREEDY[Greedy Assembly<br/>Overlap Graph]
MSA --> PROG[Progressive<br/>Alignment]
HA --> PAR[Partitioning<br/>Heuristic]
MPO --> LP[scipy.linprog<br/>LP Relaxation]
GCD --> BFS[BFS<br/>Path Finding]
PTD --> NJ[Neighbor-Joining<br/>Approximation]
RNA --> DYN[Dynamic Programming<br/>Approximation]
end
sequenceDiagram
participant CEO as CEO/Director
participant ORCH as Orchestrator
participant CL as Cluster Leader
participant W as Worker Agent
CEO->>ORCH: Strategic goal
ORCH->>CL: Decompose into clusters
CL->>W: Assign specialized tasks
W->>W: Execute (TDD)
W->>CL: Return results
CL->>ORCH: Aggregate cluster output
ORCH->>CEO: Synthesize + decide
CEO->>ORCH: Next wave / refine
graph LR
subgraph Sim["Digital Twin Simulation"]
S1[Cell Count] --> S2[Temperature]
S2 --> S3[pH]
S3 --> S4[Metabolites]
S4 --> S1
end
subgraph Real["Real World"]
R1[Experiment] --> R2[Measurement]
R2 --> R3[Analysis]
end
Sim -->|Predict| Real
Real -->|Calibrate| Sim
See docs/archify-macro-architecture.html for the full Archify-generated macro architecture diagram.
See docs/code-wiki/ for the auto-generated code wiki with Mermaid diagrams for all modules.
# Install
pip install -e ".[dev]"
# Run tests
pytest tests/ -v
# Run examples
python examples/basic_usage.py
# Docker
docker-compose up -d| Module | Description | Key Algorithms |
|---|---|---|
crispr |
Guide RNA design, off-target prediction | Rule Set 2, CFD scoring, Cas9/Cas12a/Cas13, Base/Prime Editing |
protein |
Protein sequence analysis, design | Stability prediction, pI calculation, secondary structure |
genomics |
Variant calling, assembly, phasing | Pileup calling, greedy assembly, chromosome-aware phasing |
synbio |
Genetic circuit design, pathway optimization | Logic gates, flux optimization, MoClo, RBS, promoters, terminators |
genetherapy |
Vector design, delivery optimization | Titer estimation, immunogenicity, MOI, dose optimization |
metabolic |
FBA, strain optimization | Greedy LP, FVA, GPR rules, robustness analysis, flux sampling |
aiml |
Protein LMs, variant effect, DTI | k-mer embeddings, cosine similarity, model registry, caching |
biosecurity |
Sequence screening, compliance | Risk scoring, dual-use detection, BSL classification, watermarking |
integration |
Knowledge graph, event bus, digital twin | SPARQL, GraphML, shortest path, connected components |
pipeline |
Multi-step workflow composition | Chaining, async, parallel, conditional, retry, metrics |
sizing |
Tier recommendations, cost/timeline estimates | Lab/startup/pharma profiles, serialization |
Built on 100-agent research across 10 clusters:
- CRISPR-Cas Systems (Cas9, Cas12, Cas13, base/prime editing)
- Protein Engineering (de novo design, structure prediction, stability)
- Synthetic Biology (genetic circuits, pathways, DNA synthesis)
- Gene Therapy (AAV, lentivirus, delivery, immune response)
- Computational Genomics (variants, assembly, haplotypes, MSA)
- Metabolic Engineering (FBA, strain optimization, GEMs)
- Single-Cell Genomics (scRNA-seq, spatial, multi-omics)
- Epigenomics (methylation, histones, chromatin, 3D genome)
- AI/ML for Genomics (protein LMs, variant effect, DTI, GFMs)
- Integrated Systems (KG, event buses, multi-agent, digital twins)
- Protein folding (Unger & Moult 1993)
- Protein design (Pierce & Winfree 2002)
- Genome assembly (Medvedev 2007)
- Multiple sequence alignment (Wang & Jiang 1994)
- Haplotype assembly (MEC formulation)
- Metabolic pathway optimization (MILP)
- Gene circuit design (combinatorial)
- Phylogenetic tree reconstruction
- RNA inverse folding
1114 tests passing across 11 modules. All modules follow TDD (test-first, red-green-refactor).
MIT