---
dataset_info:
  features:
  - name: seq
    dtype: string
  - name: label
    dtype: string
  splits:
  - name: train
    num_bytes: 1800054
    num_examples: 6289
  - name: valid
    num_bytes: 200145
    num_examples: 699
  - name: test
    num_bytes: 499393
    num_examples: 1745
  download_size: 299940
  dataset_size: 2499592
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: valid
    path: data/valid-*
  - split: test
    path: data/test-*
license: apache-2.0
task_categories:
- text-classification
tags:
- chemistry
- biology
- medical
size_categories:
- 1K<n<10K
---


# Dataset Card for Fitness Prediction (GB1) Dataset

### Dataset Summary

The Fitness Prediction (GB1) task is dedicated to anticipating the fitness landscape of the GB1 domain, a process that plays a pivotal role in understanding and enhancing the binding affinity and stability of this domain. As a prevalent protein interaction domain, GB1 finds wide usage in diverse applications such as protein purification, biosensors, and drug delivery. This task is configured as a regression problem, where the goal is to predict the fitness score of GB1 binding following mutations at four specific positions. 


## Dataset Structure

### Data Instances
For each instance, there is a string representing the protein sequence and a float value indicating the fitness score of the protein sequence.  See the [fitness prediction dataset viewer](https://huggingface.co/datasets/Bo1015/fitness_prediction/viewer) to explore more examples.

```
{'seq':'MEHVIDNFDNIDKCLKCGKPIKVVKLKYIKKKIENIPNSHLINFKYCSKCKRENVIENL'
'label':3.6}
```

The average  for the `seq` and the `label` are provided below:

| Feature    | Mean Count |
| ---------- | ---------------- |
| seq    |    265   |
| label  |    1.12  |




### Data Fields

- `seq`: a string containing the protein sequence
- `label`: a float value indicating the fitness score of the protein sequence.

### Data Splits

The fitness prediction dataset has 3 splits: _train_, _valid_ and _test_. Below are the statistics of the dataset.

| Dataset Split | Number of Instances in Split                |
| ------------- | ------------------------------------------- |
| Train         | 6,289                |
| Valid         | 699               |
| Test          | 1,745                           |

### Source Data

#### Initial Data Collection and Normalization
The data for this task is sourced from the [FLIP database](https://paperswithcode.com/dataset/flip).

### Licensing Information

The dataset is released under the [Apache-2.0 License](http://www.apache.org/licenses/LICENSE-2.0). 

### Citation
If you find our work useful, please consider citing the following paper:

```
@misc{chen2024xtrimopglm,
  title={xTrimoPGLM: unified 100B-scale pre-trained transformer for deciphering the language of protein},
  author={Chen, Bo and Cheng, Xingyi and Li, Pan and Geng, Yangli-ao and Gong, Jing and Li, Shen and Bei, Zhilei and Tan, Xu and Wang, Boyan and Zeng, Xin and others},
  year={2024},
  eprint={2401.06199},
  archivePrefix={arXiv},
  primaryClass={cs.CL},
  note={arXiv preprint arXiv:2401.06199}
}
```