DRAM prepare_databases Stuck at “Populating the description db” Step

[mengyuan-JI]

Description of the bug

I’m encountering a persistent issue while running DRAM-setup.py prepare_databases. The process consistently gets stuck at the final step:
Populating the description db, this may take some time
This happens even when using a high-memory node (1TB RAM)

System information

hpc

Command used and terminal output

Relevant files

No response

[Sneakypeat]

Did you find a solution? I am stuck on it. I have tried DRAM_hacks, manual DB curation, and after that, I am stuck at a key error. I have tried updating, but it is taking a long time.

[mengyuan-JI]

just skip uniprot it will work

…

On Mon, Jul 14, 2025 at 02:54 Sabih ***@***.***> wrote: *Sneakypeat* left a comment (WrightonLabCSU/DRAM#432) <#432 (comment)> Did you find any solution? I am stuck at it, I have tried DRAM_hacks, manual db curation and after that I am stuck at key error, tried update and it even it is taking long. — Reply to this email directly, view it on GitHub <#432 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ATCUT32Q6XYNFBGIUWX3NH33INOZZAVCNFSM6AAAAACAQZVR7KVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTANRYGIZDGNZQGY> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[Sneakypeat]

Just skip uniprot it will work
…

Is there a workaround if we want to include Uniprot annotations, like using a pre-built database? Or maybe wait for DRAM v2 so they fix it.

[Biocomputationalist]

To optimize the "Populating the description db" step, I recommend using an SSD. This step involves heavy read and write operations, and relying on RAM or slower storage (e.g., HDD) can significantly increase processing time. Using an SSD can reduce I/O bottlenecks and speed up the process. It worked for me along with DRAM_hacks.

[madeline-scyphers]

We have moved development to DRAM v2 now. Which you can find on dev branch. It is still a work in progress, but it is in a usable state. With DRAM v2, the databases, other than KEGG, we have pre formated and you can just download them and use without a prepare_databases step.