marcellszi / rna3db

A dataset for training and benchmarking deep learning models for RNA structure prediction
MIT License
33 stars 3 forks source link

wrong .cif file #6

Closed WangHuiNEU closed 2 months ago

WangHuiNEU commented 3 months ago

mmcif/test_set/component_98/7mq8_L0/7mq9_L0.cif, this .cif file is None, so the number of test set is 1343, not 1344 for the paper reported.

marcellszi commented 3 months ago

Thanks for finding the issue with the mmCIF file for 7mq9_L0. It appears that this file was generated incorrectly. Please give me a few days investigate.

Here is what the chain should look like according to the PDB entry: 7MQ9

Please note that rna3db-mmcifs.tar.xz is currently experimental, so there may be issues. The JSON release (rna3db-jsons.tar.gz) does not have this issue, so the numbers reported in the paper are correct.

mmagnus commented 2 months ago

@WangHuiNEU This is an interesting case. When I re-ran our pipeline, it produced a perfectly correct file. It's possible there were some I/O issues or problems during compression.

7mq9_L0.cif.zip

https://github.com/marcellszi/rna3db/releases/tag/2024-01-10