# google-deepmind/alphagenome_research architecture

> Architecture diagram of the GitHub repository google-deepmind/alphagenome_research, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-10-04). It is an AI-made overview: check details against the source.

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

## Overview

AlphaGenome Research provides a DNA sequence prediction model and a higher-level DNA Model interface for predictions and variant analysis, alongside TFRecord data loading, fine-tuning, and evaluation utilities. The main workflow is user input to DNA Model, model inference, then predictions or variant scores; documented weight providers supply pretrained checkpoints. Internal model components and data/evaluation branches are included. Relationships are limited to README statements and the supplied source index; several subsystem details were not excerpted.

Main entry points: [`src/alphagenome_research/model/dna_model.py`](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/dna_model.py) (DNA Model API), [`src/alphagenome_research/model/model.py`](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/model.py) (AlphaGenome model), [`src/alphagenome_research/evals/track_prediction.py`](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/evals/track_prediction.py) (Track evaluation), [`src/alphagenome_research/model/variant_scoring/variant_scoring.py`](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/variant_scoring/variant_scoring.py) (Variant scorers)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_interface["Model interface"]
  node_dna_model["DNA Model API<br/>[dna_model.py]"]
  node_predictions["Sequence predictions<br/>[dna_model.py]"]
end

subgraph group_scoring["Variant analysis"]
  node_variant_scoring["Variant scorers<br/>[variant_scoring.py]"]
  node_splice_scoring["Splice scoring<br/>[splice_junction.py]"]
  node_gene_masks["Gene masks"]
  node_interval_scoring["Interval scoring"]
end

subgraph group_engine["Model engine"]
  node_model["AlphaGenome model<br/>[model.py]"]
  node_heads["Prediction heads<br/>[heads.py]"]
  node_attention["Attention layers<br/>[attention.py]"]
  node_embeddings["Embeddings<br/>[embeddings.py]"]
  node_encoder["DNA encoder<br/>[one_hot_encoder.py]"]
  node_metadata["Track metadata<br/>[metadata.py]"]
end

subgraph group_data["Data and training"]
  node_tfrecords["TFRecord loader<br/>[dataset.py]"]
  node_bundles["Track bundles<br/>[bundles.py]"]
  node_finetune_data["Fine-tuning dataset<br/>[dataset.py]"]
  node_finetune["Fine-tuning<br/>[finetune.py]"]
end

subgraph group_evaluation["Evaluation"]
  node_track_eval["Track evaluation"]
  node_metrics["Regression metrics"]
end

node_user(("Research user"))
node_kaggle[("Kaggle weights")]
node_huggingface[("Hugging Face weights")]

node_user -->|"requests predictions"| node_dna_model
node_dna_model -.->|"downloads weights"| node_kaggle
node_dna_model -.->|"downloads weights"| node_huggingface
node_dna_model -->|"runs inference"| node_model
node_model -->|"produces tracks"| node_predictions
node_dna_model -->|"scores variants"| node_variant_scoring
node_dna_model -.->|"scores intervals"| node_interval_scoring
node_variant_scoring -->|"uses scorer"| node_splice_scoring
node_variant_scoring -->|"extracts masks"| node_gene_masks
node_dna_model -->|"selects track outputs"| node_metadata
node_model -->|"dispatches heads"| node_heads
node_model -->|"uses attention"| node_attention
node_model -->|"uses embeddings"| node_embeddings
node_dna_model -->|"encodes sequences"| node_encoder
node_tfrecords -->|"parses bundle types"| node_bundles
node_finetune -->|"loads training data"| node_finetune_data
node_finetune -->|"trains model"| node_model
node_track_eval -->|"loads evaluation data"| node_tfrecords
node_track_eval -->|"evaluates predictions"| node_model
node_track_eval -->|"computes metrics"| node_metrics

click node_dna_model "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/dna_model.py"
click node_predictions "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/dna_model.py"
click node_variant_scoring "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/variant_scoring/variant_scoring.py"
click node_splice_scoring "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/variant_scoring/splice_junction.py"
click node_gene_masks "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/variant_scoring/gene_mask_extractor.py"
click node_interval_scoring "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/interval_scoring/interval_scoring.py"
click node_model "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/model.py"
click node_heads "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/heads.py"
click node_attention "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/attention.py"
click node_embeddings "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/embeddings.py"
click node_encoder "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/one_hot_encoder.py"
click node_metadata "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/metadata/metadata.py"
click node_tfrecords "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/io/dataset.py"
click node_bundles "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/io/bundles.py"
click node_finetune_data "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/finetuning/dataset.py"
click node_finetune "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/finetuning/finetune.py"
click node_track_eval "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/evals/track_prediction.py"
click node_metrics "https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/evals/regression_metrics.py"

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_dna_model,node_predictions,node_user toneBlue
class node_variant_scoring,node_splice_scoring,node_gene_masks,node_interval_scoring,node_kaggle,node_huggingface toneAmber
class node_model,node_heads,node_attention,node_embeddings,node_encoder,node_metadata toneMint
class node_tfrecords,node_bundles,node_finetune_data,node_finetune toneRose
class node_track_eval,node_metrics toneIndigo
```

## Components

### Model interface

- [DNA Model API](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/dna_model.py): `src/alphagenome_research/model/dna_model.py`
- [Sequence predictions](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/dna_model.py): `src/alphagenome_research/model/dna_model.py`

### Variant analysis

- [Variant scorers](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/variant_scoring/variant_scoring.py): `src/alphagenome_research/model/variant_scoring/variant_scoring.py`
- [Splice scoring](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/variant_scoring/splice_junction.py): `src/alphagenome_research/model/variant_scoring/splice_junction.py`
- [Gene masks](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/variant_scoring/gene_mask_extractor.py): `src/alphagenome_research/model/variant_scoring/gene_mask_extractor.py`
- [Interval scoring](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/interval_scoring/interval_scoring.py): `src/alphagenome_research/model/interval_scoring/interval_scoring.py`

### Model engine

- [AlphaGenome model](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/model.py): `src/alphagenome_research/model/model.py`
- [Prediction heads](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/heads.py): `src/alphagenome_research/model/heads.py`
- [Attention layers](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/attention.py): `src/alphagenome_research/model/attention.py`
- [Embeddings](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/embeddings.py): `src/alphagenome_research/model/embeddings.py`
- [DNA encoder](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/one_hot_encoder.py): `src/alphagenome_research/model/one_hot_encoder.py`
- [Track metadata](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/model/metadata/metadata.py): `src/alphagenome_research/model/metadata/metadata.py`

### Data and training

- [TFRecord loader](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/io/dataset.py): `src/alphagenome_research/io/dataset.py`
- [Track bundles](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/io/bundles.py): `src/alphagenome_research/io/bundles.py`
- [Fine-tuning dataset](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/finetuning/dataset.py): `src/alphagenome_research/finetuning/dataset.py`
- [Fine-tuning](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/finetuning/finetune.py): `src/alphagenome_research/finetuning/finetune.py`

### Evaluation

- [Track evaluation](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/evals/track_prediction.py): `src/alphagenome_research/evals/track_prediction.py`
- [Regression metrics](https://github.com/google-deepmind/alphagenome_research/blob/main/src/alphagenome_research/evals/regression_metrics.py): `src/alphagenome_research/evals/regression_metrics.py`

### Other components

- **Research user**
- **Kaggle weights**
- **Hugging Face weights**

## Connections

- Research user → DNA Model API: requests predictions (evidence: `README.md`)
- DNA Model API → Kaggle weights: downloads weights
- DNA Model API → Hugging Face weights: downloads weights
- DNA Model API → AlphaGenome model: runs inference (evidence: `src/alphagenome_research/model/dna_model.py`)
- AlphaGenome model → Sequence predictions: produces tracks (evidence: `README.md`)
- DNA Model API → Variant scorers: scores variants (evidence: `src/alphagenome_research/model/dna_model.py`)
- DNA Model API → Interval scoring: scores intervals
- Variant scorers → Splice scoring: uses scorer (evidence: `src/alphagenome_research/model/variant_scoring/splice_junction.py`)
- Variant scorers → Gene masks: extracts masks (evidence: `src/alphagenome_research/model/variant_scoring/gene_mask_extractor.py`)
- DNA Model API → Track metadata: selects track outputs (evidence: `src/alphagenome_research/model/dna_model.py`)
- AlphaGenome model → Prediction heads: dispatches heads (evidence: `src/alphagenome_research/model/model.py`)
- AlphaGenome model → Attention layers: uses attention (evidence: `src/alphagenome_research/model/model.py`)
- AlphaGenome model → Embeddings: uses embeddings (evidence: `src/alphagenome_research/model/model.py`)
- DNA Model API → DNA encoder: encodes sequences (evidence: `src/alphagenome_research/model/dna_model.py`)
- TFRecord loader → Track bundles: parses bundle types (evidence: `src/alphagenome_research/io/dataset.py`)
- Fine-tuning → Fine-tuning dataset: loads training data (evidence: `src/alphagenome_research/finetuning/finetune.py`)
- Fine-tuning → AlphaGenome model: trains model (evidence: `src/alphagenome_research/finetuning/finetune.py`)
- Track evaluation → TFRecord loader: loads evaluation data (evidence: `src/alphagenome_research/evals/track_prediction.py`)
- Track evaluation → AlphaGenome model: evaluates predictions (evidence: `src/alphagenome_research/evals/track_prediction.py`)
- Track evaluation → Regression metrics: computes metrics (evidence: `src/alphagenome_research/evals/track_prediction.py`)
