The database generated with ProstT5 based on fasta files is incomplete

[chengaoxiang1985]

The database generated with ProstT5 based on fasta files is not complete, at least it does not contain the "XXX_ca" file。If the database is built with the pdb format files as input, there is no problem, the complete database will be generated, including the "XXX_ca" file.A complete database is required, otherwise the structure search will result in an error message that the file "XXX_ca" cannot be found.

[chengaoxiang1985]

This is my code to creat the database:

foldseek createdb query_data_cgx/QUERY.fasta database_model_cgx/my_database/DB_cgx --prostt5-model database_model_cgx/prostt5-f16-safetensors.

Followging are all the files included in the newly created database:

Followging are all the files included in the database 'pdb100' downloaded from this Github, clearly some XXX_ca files are contained:

[chengaoxiang1985]

找到答案了： https://github.com/steineggerlab/foldseek/issues/305

[xjhzjucas]

The database generated with ProstT5 based on fasta files is not complete, at least it does not contain the "XXX_ca" file。If the database is built with the pdb format files as input, there is no problem, the complete database will be generated, including the "XXX_ca" file.A complete database is required, otherwise the structure search will result in an error message that the file "XXX_ca" cannot be found.

您好，想问一下您可以通过foldseek databases命令成功下载各个数据库吗？好像有一些墙的问题，不知道能否解决，谢谢！

[chengaoxiang1985]

这个问题我们也有，也是翻墙下载后手动上传过去的。

[xjhzjucas]

好的感谢！有一个疑问是直接手动上传后，不知道需不需要建索引等其他操作呀？因为感觉foldseek databases命令应该不仅仅是下载文件，不知道您有没有经验

[chengaoxiang1985]

可以解压放在任何地方，只要在用的时候放上绝对路径，如果在当前路径下，则以当前路径开头补上下载到的文件的路径即可，完全不影响结果。