How to index a large fasta file (>100GB) by using bowtie2-build

0

Entering edit mode

4.7 years ago

taylor • 0

I am wondering how to index a very large fasta file (~100GB) by using bowtie2-build? I have enlarged my memory up to 512G and use the parameter "--bmax 3 --dcv 4096" to minimize the memory usage, but it still failed...

So is there any possible solution to solve this problem? Thanks~

Here is the error message:

Using parameters --bmax 3 --dcv 4096

Doing ahead-of-time memory usage test

Ran out of memory; automatically trying more memory-economical parameters.

Total time for call to driver() for forward index: 01:17:56

Thanks,

alignment genome bowtie index • 2.4k views

ADD COMMENT • link updated 4.7 years ago by WouterDeCoster 47k • written 4.7 years ago by taylor • 0

1

Entering edit mode

Centrifuge also can be useful https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5131823/

ADD REPLY • link 4.7 years ago by grant.hovhannisyan ★ 2.6k

0

Entering edit mode

Are you sure bowtie is the right tool for the job? Which index are you building?

ADD REPLY • link 4.7 years ago by WouterDeCoster 47k

0

Entering edit mode

They are some environment microbes... I think bowtie is easy to use. Do you have some software recommended for this job? Thanks,

ADD REPLY • link 4.7 years ago by taylor • 0

1

Entering edit mode

Those microbes may have highly similar sequences, making alignment erroneous. Take a look at kraken2 for species classification

ADD REPLY • link 4.7 years ago by WouterDeCoster 47k

0

Entering edit mode

OK. Thanks. I will try.

ADD REPLY • link 4.7 years ago by taylor • 0

0

Entering edit mode

Maybe you could try bwa mem ? what do you want to do with the index ? which type of data ? which type of organism are you studying ?

ADD REPLY • link 4.7 years ago by Nicolas Rosewick 10k

Login before adding your answer.