Could you help me ? I'm confused with velvet which need a large physical memory.

[jzhou65]

$ velvetg output -exp_cov auto
[0.000001] Reading read set file output/Sequences;
[43.817688] 88015946 sequences found
[110.015839] Done
[218.490050] Reading pre-graph file output/PreGraph
[218.490744] Graph has 19624790 nodes and 88015946 sequences
[233.340736] Scanning pre-graph file output/PreGraph for k-mers
[236.788591] 123729624 kmers found
[250.566077] Sorting kmer occurence table ...
[335.348217] Sorting done.
[335.348450] Computing acceleration table...
[336.150998] Computing offsets...
[337.488936] Ghost Threading through reads 0 / 88015946

this required too large physical memory, I would like to know that is this right? my raw data is 28G.

[dzerbino]

Dear @jzhou65 ,

It's hard to tell without more information, but Velvet can be rather greedy with memory. How much memory do you have?

Best regards,

Daniel

[jzhou65]

Dear @jzhou65 ,

It's hard to tell without more information, but Velvet can be rather greedy with memory. How much memory do you have?

Best regards,

Daniel

hi, Daniel. I only used two codes as follows:

velveth output/ 17
-shortPaired -fastq -separate HJY_2020_Clean_Data1.fq HJY_2020_Clean_Data2.fq

velvetg output -exp_cov auto -cov_cutoff auto
-shortMatePaired3 yes -shortMatePaired4 yes
-clean yes -scaffolding yes -amos_file yes

the server has 384G physical memory and I used swap to add 384G Virtual memory, but the second code may be require larger memory, so I don't know how to treat with it. Thanks a lot.

[dzerbino]

Dear @jzhou65 ,

Could you please tell me more about your data? What is the size of your genome and the coverage?

Regards,

Daniel