---
license: mit
configs:
- config_name: dataset
  data_files: "dataset.csv"
---

# Description
Structure Similarity Prediction predicts the (aligned) Local Distance Difference Test (LDDT) of the structures given an unaligned pair of proteins. Target values are computed after alignment with TM-align for all pairs of 1000 randomly sampled single-chain proteins.

# Splits

**Structure type:** PDB

The dataset is from [**ProteinShake Building datasets and benchmarks for deep learning on protein structures**](https://papers.nips.cc/paper_files/paper/2023/file/b6167294ed3d6fc61e11e1592ce5cb77-Paper-Datasets_and_Benchmarks.pdf). We use the splits based on 70% structure similarity, with the number of training, validation and test set shown below:

- Train: 300699
- Valid: 4559
- Test:  4850

# Label

The label means the (aligned) Local Distance Difference Test (LDDT) of the structures given an unaligned pair of proteins. Target values are computed after alignment with TM-align for all pairs of 1000 randomly sampled single-chain proteins. This value ranges from 0 to 1.