# google-deepmind/alphaprotein-novo architecture

> Architecture diagram of the GitHub repository google-deepmind/alphaprotein-novo, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-06). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/google-deepmind/alphaprotein-novo
- Repository: https://github.com/google-deepmind/alphaprotein-novo

## Overview

AlphaProtein Novo designs enzymes by generating protein structure and sequence conditioned on a catalytic motif and ligand, then optionally redesigning sequence, predicting structures, and evaluating designs. The documented entry point is a manifest-driven pipeline that runs these stages and can resume campaigns. The graph includes motif/data preparation and the model’s sampling layer as internal components. Stage links follow the README; source excerpts are partial, so internal links are limited to relationships supported by the SOURCE INDEX or excerpts.

Main entry points: [`run_pipeline.py`](https://github.com/google-deepmind/alphaprotein-novo/blob/main/run_pipeline.py) (Pipeline runner), [`evaluate_design.py`](https://github.com/google-deepmind/alphaprotein-novo/blob/main/evaluate_design.py) (Metrics runner), [`run_generator.py`](https://github.com/google-deepmind/alphaprotein-novo/blob/main/run_generator.py) (Design generator), [`src/alphaprotein_novo/data/design_manifest.py`](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/data/design_manifest.py) (Manifest and design jobs)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_orchestration["Campaign orchestration"]
  node_pipeline["Pipeline runner<br/>[run_pipeline.py]"]
  node_outputs["Design artifacts and reports"]
end

subgraph group_generation["Motif-conditioned generation"]
  node_generator["Design generator<br/>[run_generator.py]"]
  node_motif["Motif specification<br/>[motif_spec.py]"]
  node_sampling["Diffusion sampling<br/>[sampling.py]"]
  node_denoiser["Protein denoiser<br/>[denoiser.py]"]
  node_modelconfig["Model configuration<br/>[config.py]"]
  node_modelparams["Generator parameters<br/>[params.py]"]
end

subgraph group_structure["Structure and sequence stages"]
  node_ligandmpnn["Sequence redesign<br/>[run_ligandmpnn.py]"]
  node_af3["Structure prediction<br/>[run_alphafold.py]"]
end

subgraph group_evaluation["Design evaluation"]
  node_evaluator["Metrics runner<br/>[evaluate_design.py]"]
  node_metrics["Structure and enzyme metrics"]
  node_ligmetrics["Ligand metrics<br/>[ligand_metrics.py]"]
  node_enzyme["Enzyme-specific metrics"]
end

subgraph group_foundation["Shared modeling data"]
  node_manifest["Manifest and design jobs<br/>[design_manifest.py]"]
  node_features["Structure features"]
  node_structutils["Structure utilities<br/>[structure_utils.py]"]
  node_liganddata["Ligand data<br/>[ligand_data.py]"]
end

node_user(("Design engineer"))
node_af3service["AlphaFold 3"]

node_user -->|"configures campaign"| node_pipeline
node_pipeline -->|"loads settings"| node_manifest
node_pipeline -->|"runs generation"| node_generator
node_pipeline -.->|"optionally redesigns"| node_ligandmpnn
node_pipeline -->|"runs prediction"| node_af3
node_pipeline -->|"runs evaluation"| node_evaluator
node_generator -->|"reads jobs"| node_manifest
node_generator -->|"uses motif specs"| node_motif
node_generator -->|"uses sampler"| node_sampling
node_sampling -->|"calls denoiser"| node_denoiser
node_denoiser -->|"uses configuration"| node_modelconfig
node_generator -->|"loads parameters"| node_modelparams
node_motif -->|"uses feature utilities"| node_features
node_motif -->|"uses structure helpers"| node_structutils
node_structutils -->|"uses ligand data"| node_liganddata
node_ligandmpnn -->|"updates structures"| node_structutils
node_af3 -->|"reads prediction specs"| node_manifest
node_af3 -->|"invokes folding"| node_af3service
node_evaluator -->|"loads design settings"| node_manifest
node_evaluator -->|"computes metrics"| node_metrics
node_evaluator -->|"computes ligand scores"| node_ligmetrics
node_evaluator -->|"selects evaluation suite"| node_enzyme
node_metrics -->|"uses ligand metrics"| node_ligmetrics
node_evaluator -.->|"writes reports"| node_outputs
node_pipeline -->|"organizes stage artifacts"| node_outputs

click node_manifest "https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/data/design_manifest.py"
click node_pipeline "https://github.com/google-deepmind/alphaprotein-novo/blob/main/run_pipeline.py"
click node_generator "https://github.com/google-deepmind/alphaprotein-novo/blob/main/run_generator.py"
click node_motif "https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/data/motif_spec.py"
click node_sampling "https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/model/sampling.py"
click node_denoiser "https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/model/denoiser.py"
click node_modelconfig "https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/model/config.py"
click node_modelparams "https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/model/params.py"
click node_features "https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/data/structure_features.py"
click node_structutils "https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/data/structure_utils.py"
click node_liganddata "https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/data/ligand_data.py"
click node_ligandmpnn "https://github.com/google-deepmind/alphaprotein-novo/blob/main/run_ligandmpnn.py"
click node_af3 "https://github.com/google-deepmind/alphaprotein-novo/blob/main/run_alphafold.py"
click node_evaluator "https://github.com/google-deepmind/alphaprotein-novo/blob/main/evaluate_design.py"
click node_metrics "https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/metrics/structure_metrics.py"
click node_ligmetrics "https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/metrics/ligand_metrics.py"
click node_enzyme "https://github.com/google-deepmind/alphaprotein-novo/tree/main/src/alphaprotein_novo/metrics"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_pipeline,node_outputs toneBlue
class node_generator,node_motif,node_sampling,node_denoiser,node_modelconfig,node_modelparams toneAmber
class node_ligandmpnn,node_af3 toneMint
class node_evaluator,node_metrics,node_ligmetrics,node_enzyme toneRose
class node_manifest,node_features,node_structutils,node_liganddata,node_user,node_af3service toneIndigo
```

## Components

### Campaign orchestration

- [Pipeline runner](https://github.com/google-deepmind/alphaprotein-novo/blob/main/run_pipeline.py): `run_pipeline.py`
- **Design artifacts and reports**

### Motif-conditioned generation

- [Design generator](https://github.com/google-deepmind/alphaprotein-novo/blob/main/run_generator.py): `run_generator.py`
- [Motif specification](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/data/motif_spec.py): `src/alphaprotein_novo/data/motif_spec.py`
- [Diffusion sampling](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/model/sampling.py): `src/alphaprotein_novo/model/sampling.py`
- [Protein denoiser](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/model/denoiser.py): `src/alphaprotein_novo/model/denoiser.py`
- [Model configuration](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/model/config.py): `src/alphaprotein_novo/model/config.py`
- [Generator parameters](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/model/params.py): `src/alphaprotein_novo/model/params.py`

### Structure and sequence stages

- [Sequence redesign](https://github.com/google-deepmind/alphaprotein-novo/blob/main/run_ligandmpnn.py): `run_ligandmpnn.py`
- [Structure prediction](https://github.com/google-deepmind/alphaprotein-novo/blob/main/run_alphafold.py): `run_alphafold.py`

### Design evaluation

- [Metrics runner](https://github.com/google-deepmind/alphaprotein-novo/blob/main/evaluate_design.py): `evaluate_design.py`
- [Structure and enzyme metrics](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/metrics/structure_metrics.py): `src/alphaprotein_novo/metrics/structure_metrics.py`
- [Ligand metrics](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/metrics/ligand_metrics.py): `src/alphaprotein_novo/metrics/ligand_metrics.py`
- [Enzyme-specific metrics](https://github.com/google-deepmind/alphaprotein-novo/tree/main/src/alphaprotein_novo/metrics): `src/alphaprotein_novo/metrics`

### Shared modeling data

- [Manifest and design jobs](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/data/design_manifest.py): `src/alphaprotein_novo/data/design_manifest.py`
- [Structure features](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/data/structure_features.py): `src/alphaprotein_novo/data/structure_features.py`
- [Structure utilities](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/data/structure_utils.py): `src/alphaprotein_novo/data/structure_utils.py`
- [Ligand data](https://github.com/google-deepmind/alphaprotein-novo/blob/main/src/alphaprotein_novo/data/ligand_data.py): `src/alphaprotein_novo/data/ligand_data.py`

### Other components

- **Design engineer**
- **AlphaFold 3**

## Connections

- Design engineer → Pipeline runner: configures campaign (evidence: `README.md`)
- Pipeline runner → Manifest and design jobs: loads settings (evidence: `README.md`)
- Pipeline runner → Design generator: runs generation (evidence: `README.md`)
- Pipeline runner → Sequence redesign: optionally redesigns (evidence: `README.md`)
- Pipeline runner → Structure prediction: runs prediction (evidence: `README.md`)
- Pipeline runner → Metrics runner: runs evaluation (evidence: `README.md`)
- Design generator → Manifest and design jobs: reads jobs (evidence: `run_generator.py`)
- Design generator → Motif specification: uses motif specs (evidence: `run_generator.py`)
- Design generator → Diffusion sampling: uses sampler (evidence: `run_generator.py`)
- Diffusion sampling → Protein denoiser: calls denoiser (evidence: `src/alphaprotein_novo/model/sampling.py`)
- Protein denoiser → Model configuration: uses configuration (evidence: `src/alphaprotein_novo/model/denoiser.py`)
- Design generator → Generator parameters: loads parameters (evidence: `run_generator.py`)
- Motif specification → Structure features: uses feature utilities (evidence: `src/alphaprotein_novo/data/motif_spec.py`)
- Motif specification → Structure utilities: uses structure helpers (evidence: `src/alphaprotein_novo/data/motif_spec.py`)
- Structure utilities → Ligand data: uses ligand data (evidence: `src/alphaprotein_novo/data/structure_utils.py`)
- Sequence redesign → Structure utilities: updates structures (evidence: `run_ligandmpnn.py`)
- Structure prediction → Manifest and design jobs: reads prediction specs (evidence: `run_alphafold.py`)
- Structure prediction → AlphaFold 3: invokes folding (evidence: `README.md`)
- Metrics runner → Manifest and design jobs: loads design settings (evidence: `evaluate_design.py`)
- Metrics runner → Structure and enzyme metrics: computes metrics (evidence: `evaluate_design.py`)
- Metrics runner → Ligand metrics: computes ligand scores (evidence: `evaluate_design.py`)
- Metrics runner → Enzyme-specific metrics: selects evaluation suite (evidence: `README.md`)
- Structure and enzyme metrics → Ligand metrics: uses ligand metrics (evidence: `src/alphaprotein_novo/metrics/structure_metrics.py`)
- Metrics runner → Design artifacts and reports: writes reports
- Pipeline runner → Design artifacts and reports: organizes stage artifacts (evidence: `README.md`)
