search

vrmarcelino / CCMetagen

Microbiome classification pipeline
GNU General Public License v3.0
64 stars 19 forks source link

310 Segmentation fault (core dumped) #46

Closed taniagmangolini closed 1 year ago

taniagmangolini commented 1 year ago

Hi, I am facing the error 310 Segmentation fault (core dumped) when running KMA. I am using the following command to process a fasta file that has 10.7 MB in size:

kma -i scaffolds.fasta -o scaffolds.kmaout -t_db ncbi_nt_no_env_11jun2019 -ca -t 1 -1t1 -mem_mode -ef

I am using kma v 1.4.6 and the machine has 500GB RAM.

Could you help me to understand and fix this problem?

Thank you in advance!

vrmarcelino commented 1 year ago

Hi!

It is possible that the temp files are being saved somewhere where there is no space. Could you try adding a temp folder with the flag '-tmp' ?

Also, it might be problematic to use contigs AND use the -1t1 AND the NCBI nt database. You will have multiple genes in each contig but you are telling KMA to match just one reference sequence (often individual genes in NCBI nt) to each contig.

If you have enough depth to assemble MAGs, I would recommend classifying them with GTDBtk (although it will miss the eukaryotes and viruses). You could combine that with a read-level CCMetagen analyses to identify the eukaryotes + viruses + rare species that did not assemble.

OR: If you are working with a well-known environment (human gut), you can build a reference database containing complete (or draft) genomes only, and then do the assignments with KMA (using the '-ca' and the '-1t1' flags)

vrmarcelino commented 1 year ago

Closing issue due to inactivity. Feel free to open it again if you need help.