File size: 3,918 Bytes
6acfefd
e074056
6acfefd
e074056
 
 
 
 
 
73fbcca
 
 
 
0779e53
73fbcca
0779e53
73fbcca
 
 
0779e53
73fbcca
 
 
3ab1c1a
73fbcca
 
 
0779e53
73fbcca
 
 
0779e53
1da983e
 
 
 
6acfefd
e074056
73fbcca
e074056
7c97de1
ccc6e4b
7c97de1
e074056
ab1ca2e
e074056
3ab1c1a
e074056
3ab1c1a
e074056
3ab1c1a
 
 
 
ab1ca2e
e074056
1da983e
e074056
73fbcca
3ab1c1a
e074056
3ab1c1a
 
1da983e
3ab1c1a
e074056
3ab1c1a
73fbcca
0779e53
e074056
3ab1c1a
e074056
484ec34
e074056
3ab1c1a
e074056
3ab1c1a
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
---
pretty_name: CryoAtlas
license: cc-by-nc-4.0
tags:
- cryo-em
- structural-biology
- protein-structure
- benchmark
configs:
- config_name: catalog
  default: true
  data_files:
  - split: train
    path: resources/metadata/catalog/train.parquet
  - split: test
    path: resources/metadata/catalog/test.parquet
- config_name: benchmark_targets
  data_files:
  - split: test
    path: resources/metadata/catalog/benchmark_targets.parquet
- config_name: benchmark_results
  data_files:
  - split: test
    path: resources/benchmark/results.parquet
- config_name: examples
  data_files:
  - split: examples
    path: resources/benchmark/examples.parquet
- config_name: downloads
  data_files:
  - split: archives
    path: resources/metadata/catalog/archives.parquet
- config_name: training_utility
  data_files:
  - split: test
    path: resources/benchmark/training_utility/results.parquet
---

# CryoAtlas

A large curated dataset and unified benchmark for cryo-EM atomic model building.

*Accepted at NeurIPS 2026 · Evaluations & Datasets Track*

[Paper (OpenReview)](https://openreview.net/forum?id=qHJrv65ncA) · [Code](https://github.com/oulling/CryoAtlas-Code) · [Project website](https://oulling.github.io/cryoatlas/) · [3D Benchmark Explorer](https://oulling.github.io/cryoatlas/explore/)

CryoAtlas v1.0 provides **17,809 map–model–sequence entries** derived from EMDB and PDB records available up to March 2026: **16,657 training entries** and **1,152 test entries**, including the **618-target common benchmark**. The benchmark evaluates structural accuracy, map agreement and stereochemical quality.

## Download

| Folder | Contents | Compressed size |
|---|---|---:|
| [**train/**](https://huggingface.co/datasets/markfaa/CryoAtlas/tree/main/train) | Four disjoint training parts, 68 independent archives | 2.264 TB |
| [**test/**](https://huggingface.co/datasets/markfaa/CryoAtlas/tree/main/test) | Two archives: 618 benchmark targets + 534 remaining test entries | 150.9 GB |
| [**resources/**](https://huggingface.co/datasets/markfaa/CryoAtlas/tree/main/resources) | Benchmark resource package, metadata, indexes and download tools | Resource package: 1.90 GB |

**Start small:** [download the benchmark resource package](https://huggingface.co/datasets/markfaa/CryoAtlas/resolve/main/resources/CryoAtlas_v1.0_benchmark_resources.tar.gz) for baseline and retrained-ModelAngelo results, predicted structures and 10 complete examples. For selected training parts or an entry's containing archive, see the [download guide](resources/docs/downloads.md).

<a id="examples"></a>
## Explore online

- [**Entry catalog**](https://huggingface.co/datasets/markfaa/CryoAtlas/viewer/catalog/train): search EMD/PDB IDs, resolution and archive links.
- [**Benchmark results**](https://huggingface.co/datasets/markfaa/CryoAtlas/viewer/benchmark_results/test): 618 targets × five methods.
- [**Training utility**](https://huggingface.co/datasets/markfaa/CryoAtlas/viewer/training_utility/test): the two released ModelAngelo retraining conditions.
- [**Example previews**](https://huggingface.co/datasets/markfaa/CryoAtlas/viewer/examples/examples): density projections from 10 released examples.

The Viewer uses lightweight tables; maps, structures and sequences are in the archives. **The 618 benchmark targets are part of the 1,152-entry test set.**

![Dataset splits and resolution distributions](resources/docs/assets/dataset_overview.png)

## Documentation

[Croissant metadata](resources/metadata/croissant.json) · [Dataset guide](resources/docs/dataset_guide.md) · [Fields and metrics](resources/docs/data_dictionary.md) · [Archive checksums](resources/metadata/checksums/SHA256SUMS)

All 70 training/test archives have verified sizes and SHA-256 hashes. The download helper verifies selected archives automatically.

**License:** [CC BY-NC 4.0](https://creativecommons.org/licenses/by-nc/4.0/).