Dataset

You can download SCOPe v2.07 (40% ID filtered subset) from its website. We remove some structures due to the technical issues. The sid of remaining structures after removal is provided here.

Besides SCOPe v2.07, we also construct an independent dataset to evaluate our model and other state-of-the-art methods. This dataset consists of protein structures derived from PDB, the release date of which is from Oct 1st 2017 to Oct 1st 2019. If a certain PDB file contains multiple chains, it will be split into multiple files, each of which contains only one chain. Then, CD-HIT is used to remove redundant sequences. The sequence identity of the independent set and that between the independent set and SCOPe v2.07 are both under 40% after filtering. At last, 51 protein chains are removed due to the technical issues. The final independent test set (named ind_PDB) contains 1,914 protein structures.

You can easily download the structure files of ind_PDB from PDB website according to the PDB ID and chain ID of each structure, which can be downloaded here. For example, 5ypzA means chain A of 5YPZ.