alchemab/sarscov2-binding-prediction
收藏Hugging Face2023-09-12 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/alchemab/sarscov2-binding-prediction
下载链接
链接失效反馈官方服务：
资源简介：
---
dataset_info:
features:
- name: vh
dtype: string
- name: vl
dtype: string
- name: label
dtype:
class_label:
names:
'0': non-binder
'1': binder
splits:
- name: train
num_bytes: 10272366
num_examples: 41787
- name: eval
num_bytes: 1141623
num_examples: 4644
- name: test
num_bytes: 1268252
num_examples: 5159
download_size: 1194761
dataset_size: 12682241
---
# SARS-CoV-2 binding dataset
Dataset of 104972 antibodies screened for binding the SARS-CoV-2 HR peptide, described in [Engelhart et al. (2022)](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9606274/), were obtained from [Zenodo](https://zenodo.org/record/5095284).
Average predicted logK<sub>D</sub> values were used for classifying sequences as binders and non-binders:
* logK<sub>D</sub><3 = binders
* logK<sub>D</sub>>=4 = non-binders
* logK<sub>D</sub>>=3 and logK<sub>D</sub> = ambiguous; removed.
Using these criteria, we have 51590 sequences remaining; these were stratified into an 80:10:10 ratio for training, test, validation, leading to:
* 41787 sequences in training
* 5159 sequences in validation
* 4644 sequences in test
Example
| vh | vl | label |
| ------- | ------- | ----- |
| EVQ... | DIQ... | 1 |
| EVQ... | DIQ... | 0 |
| EVQ... | DIQ... | 0 |
| EVQ... | DIQ... | 1 |
References
* [Engelhart et al. (2022) paper](https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9606274/)
* [Zenodo link for dataset](https://zenodo.org/record/5095284)
提供机构：
alchemab
