# dauparas/proteinmpnn architecture

> Architecture diagram of the GitHub repository dauparas/proteinmpnn, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-09-26). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/dauparas/proteinmpnn
- Repository: https://github.com/dauparas/proteinmpnn

## Overview

ProteinMPNN designs protein sequences from backbone structures and also scores sequences or reports residue probabilities. The principal workflow is structure parsing, optional chain/position and sequence-bias configuration, model inference using selected weights, then sequence or score/probability output. The README documents these capabilities; sampled source confirms the runner, parser, model components, and several constraint helpers. Some runner internals and optional modes were not sampled, so uncertain implementation links are omitted.

Main entry points: [`protein_mpnn_run.py`](https://github.com/dauparas/proteinmpnn/blob/main/protein_mpnn_run.py) (Inference runner), [`training/training.py`](https://github.com/dauparas/proteinmpnn/blob/main/training/training.py) (Training loop), [`helper_scripts/parse_multiple_chains.py`](https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/parse_multiple_chains.py) (PDB parser), [`helper_scripts/assign_fixed_chains.py`](https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/assign_fixed_chains.py) (Chain assignment)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_prep["Structure preparation"]
  node_parser["PDB parser"]
  node_parsed["Parsed structure"]
end

subgraph group_config["Design constraints"]
  node_chain_config["Chain assignment"]
  node_fixed_config["Fixed positions"]
  node_tied_config["Tied positions"]
  node_bias_config["AA bias<br/>[make_bias_AA.py]"]
  node_pssm_config["PSSM input"]
end

subgraph group_infer["Inference and outputs"]
  node_runner["Inference runner"]
  node_utilities["Parsing and data utilities"]
  node_model["ProteinMPNN model"]
  node_weights[("Model weights")]
  node_results["Sequences and scores"]
end

subgraph group_train["Model training"]
  node_train_entry["Training loop<br/>[training.py]"]
  node_train_data["Training data utilities<br/>[utils.py]"]
  node_train_model["Training model<br/>[model_utils.py]"]
  node_trained_weights[("Trained weights")]
end

node_user(("Designer"))
node_pdb["PDB backbone"]

node_user -->|"provides structure"| node_pdb
node_pdb -->|"parsed by"| node_parser
node_parser -->|"writes JSONL"| node_parsed
node_parsed -->|"configures chains"| node_chain_config
node_parsed -->|"configures positions"| node_fixed_config
node_parsed -->|"configures ties"| node_tied_config
node_parsed -->|"configures bias"| node_bias_config
node_parsed -->|"configures PSSM"| node_pssm_config
node_user -->|"runs design or scoring"| node_runner
node_runner -->|"uses parsing utilities"| node_utilities
node_runner -->|"runs inference"| node_model
node_runner -->|"loads weights"| node_weights
node_runner -->|"reads structure data"| node_parsed
node_runner -->|"reads chain settings"| node_chain_config
node_runner -->|"reads fixed positions"| node_fixed_config
node_runner -->|"reads tied positions"| node_tied_config
node_runner -->|"reads amino-acid bias"| node_bias_config
node_runner -->|"reads PSSM settings"| node_pssm_config
node_runner -->|"writes sequences and scores"| node_results
node_train_entry -->|"loads training data"| node_train_data
node_train_entry -->|"trains model"| node_train_model
node_train_entry -->|"saves checkpoints"| node_trained_weights

click node_parser "https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/parse_multiple_chains.py"
click node_chain_config "https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/assign_fixed_chains.py"
click node_fixed_config "https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/make_fixed_positions_dict.py"
click node_tied_config "https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/make_tied_positions_dict.py"
click node_bias_config "https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/make_bias_AA.py"
click node_pssm_config "https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/make_pssm_input_dict.py"
click node_runner "https://github.com/dauparas/proteinmpnn/blob/main/protein_mpnn_run.py"
click node_utilities "https://github.com/dauparas/proteinmpnn/blob/main/protein_mpnn_utils.py"
click node_model "https://github.com/dauparas/proteinmpnn/blob/main/protein_mpnn_utils.py"
click node_weights "https://github.com/dauparas/proteinmpnn/tree/main/vanilla_model_weights"
click node_results "https://github.com/dauparas/proteinmpnn/tree/main/outputs"
click node_train_entry "https://github.com/dauparas/proteinmpnn/blob/main/training/training.py"
click node_train_data "https://github.com/dauparas/proteinmpnn/blob/main/training/utils.py"
click node_train_model "https://github.com/dauparas/proteinmpnn/blob/main/training/model_utils.py"
click node_trained_weights "https://github.com/dauparas/proteinmpnn/tree/main/training/exp_020/model_weights"

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_parser,node_parsed toneBlue
class node_chain_config,node_fixed_config,node_tied_config,node_bias_config,node_pssm_config toneAmber
class node_runner,node_utilities,node_model,node_weights,node_results toneMint
class node_train_entry,node_train_data,node_train_model,node_trained_weights toneRose
class node_user,node_pdb toneIndigo
```

## Components

### Structure preparation

- [PDB parser](https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/parse_multiple_chains.py): `helper_scripts/parse_multiple_chains.py`
- **Parsed structure**

### Design constraints

- [Chain assignment](https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/assign_fixed_chains.py): `helper_scripts/assign_fixed_chains.py`
- [Fixed positions](https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/make_fixed_positions_dict.py): `helper_scripts/make_fixed_positions_dict.py`
- [Tied positions](https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/make_tied_positions_dict.py): `helper_scripts/make_tied_positions_dict.py`
- [AA bias](https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/make_bias_AA.py): `helper_scripts/make_bias_AA.py`
- [PSSM input](https://github.com/dauparas/proteinmpnn/blob/main/helper_scripts/make_pssm_input_dict.py): `helper_scripts/make_pssm_input_dict.py`

### Inference and outputs

- [Inference runner](https://github.com/dauparas/proteinmpnn/blob/main/protein_mpnn_run.py): `protein_mpnn_run.py`
- [Parsing and data utilities](https://github.com/dauparas/proteinmpnn/blob/main/protein_mpnn_utils.py): `protein_mpnn_utils.py`
- [ProteinMPNN model](https://github.com/dauparas/proteinmpnn/blob/main/protein_mpnn_utils.py): `protein_mpnn_utils.py`
- [Model weights](https://github.com/dauparas/proteinmpnn/tree/main/vanilla_model_weights): `vanilla_model_weights`
- [Sequences and scores](https://github.com/dauparas/proteinmpnn/tree/main/outputs): `outputs`

### Model training

- [Training loop](https://github.com/dauparas/proteinmpnn/blob/main/training/training.py): `training/training.py`
- [Training data utilities](https://github.com/dauparas/proteinmpnn/blob/main/training/utils.py): `training/utils.py`
- [Training model](https://github.com/dauparas/proteinmpnn/blob/main/training/model_utils.py): `training/model_utils.py`
- [Trained weights](https://github.com/dauparas/proteinmpnn/tree/main/training/exp_020/model_weights): `training/exp_020/model_weights`

### Other components

- **Designer**
- **PDB backbone**

## Connections

- Designer → PDB backbone: provides structure
- PDB backbone → PDB parser: parsed by
- PDB parser → Parsed structure: writes JSONL
- Parsed structure → Chain assignment: configures chains
- Parsed structure → Fixed positions: configures positions
- Parsed structure → Tied positions: configures ties
- Parsed structure → AA bias: configures bias
- Parsed structure → PSSM input: configures PSSM
- Designer → Inference runner: runs design or scoring
- Inference runner → Parsing and data utilities: uses parsing utilities
- Inference runner → ProteinMPNN model: runs inference
- Inference runner → Model weights: loads weights
- Inference runner → Parsed structure: reads structure data
- Inference runner → Chain assignment: reads chain settings
- Inference runner → Fixed positions: reads fixed positions
- Inference runner → Tied positions: reads tied positions
- Inference runner → AA bias: reads amino-acid bias
- Inference runner → PSSM input: reads PSSM settings
- Inference runner → Sequences and scores: writes sequences and scores
- Training loop → Training data utilities: loads training data
- Training loop → Training model: trains model
- Training loop → Trained weights: saves checkpoints
