Kraken Database Size

[asommer101-dot]

Hello,

I was able to run the MEDI build pipeline without getting an error message (although I had to use the K2 functions instead of kraken2-build as I otherwise had rsync problems).

While I am able to run the quant script now and MEDI will produce an output, I believe that there could be an issue with my database size. I specified a max size of 500GB in my script; however, my final hash.k2d file was roughly 354GB and the database.kraken is 155GB.

I would imagine that this could impact the sensitivity of the results as I ran a few test samples an the only food matches in the output were Teas.

Is this an issue you have experienced or do you have any thoughts on why there is a large database size mismatch?

Thank you in advance,

[cdiener]

Sorry for not seeing this earlier. That depends a lot on when you downloaded the data and how much could be downloaded successfully. The updates from today make this much more apparent now. As a reference here are my file sizes for a recent build:

total 566G
drwxrws---  3 o_diener dienerlab  377 Nov 17  2025 .
drwxrws--- 12 o_diener dienerlab 1,4K Aug  3 13:02 ..
-rwxrwx---  1 o_diener dienerlab 5,2M Nov 17  2025 database100mers.kmer_distrib
-rwxrwx---  1 o_diener dienerlab 4,8M Nov 17  2025 database150mers.kmer_distrib
-rwxrwx---  1 o_diener dienerlab 4,1M Nov 17  2025 database250mers.kmer_distrib
-rwxrwx---  1 o_diener dienerlab 149M Nov 17  2025 food_contents.csv.gz
-rwxrwx---  1 o_diener dienerlab  90K Nov 17  2025 food_matches.csv
-rwxrwx---  1 o_diener dienerlab 496G Nov 17  2025 hash.k2d
-rwxrwx---  1 o_diener dienerlab 169K Nov 17  2025 manifest.csv
-rwxrwx---  1 o_diener dienerlab   64 Nov 17  2025 opts.k2d
-rwxrwx---  1 o_diener dienerlab 1,2G Nov 17  2025 seqid2taxid.map
-rwxrwx---  1 o_diener dienerlab 4,5M Nov 17  2025 taxo.k2d
drwxrws---  2 o_diener dienerlab  527 Nov 17  2025 taxonomy