Segmentation Fault when comparing two databases about foldseek

[bordin89]

I tried to feed the program with 1000 structures from the query database and it ran to completion. With 2000 structures it gets the same error. Could it be some missing flags or a lack of RAM on my end?

[martin-steinegger]

Thank you for reporting it. Could you please share your 2,000 structures that causes the error?

[bordin89]

Hi, Thank you for the quick reply!

By running the data in chunks I managed to figure out which ones were the problematic PDB files. It boils down to 14 AlphaFold models for 2 proteins that are huge (M0QZD8 and P04114) with 3,266 and 4,563 residues.
I'm attaching two links to OneDrive with the structures and the database
Problematic structures https://liveuclac-my.sharepoint.com/:u:/g/personal/ucbtnb4_ucl_ac_uk/ETEYfojc9DRKhPyqRVOpAXEBfw6xkHOpTVYS46zbGctD5A?e=cRqNve
Database: https://liveuclac-my.sharepoint.com/:u:/g/personal/ucbtnb4_ucl_ac_uk/EezQQjKSFppJsK3da9a6pCQB15lR6-4WSF_AZX_dSwaPug?e=RaAXsY

Is it possible to output in the log which structure in a set is breaking the run?

Thanks again!

[martin-steinegger]

@bordin89 thank you so much for sharing the data. I ran foldseek with your example but it did not crash. Could you please try the newest version. I recently fixed a createdb bug, which might have cause the issue.

[martin-steinegger]

Does this issue still occur?

[bordin89]

No, the latest release of Foldseek seems to have fixed the issue. Thanks for the follow-up!