Requirements for running HISAT2
0
0
Entering edit mode
6.4 years ago
Arindam Ghosh ▴ 510

What is the minimum computational requirements to run hisat2-build command using --ss (splicesites) and --exon (exons)? Though it has been recommended to have 200GB memory is it possible to run with lower RAM especially since it looks to using more economical parameters.

I am trying to build an index for Human genome. I have tried to execute that in a desktop computer with Intel core i5 processor and 4GB ram but the job was killed after multiple attempts to use economical parameters.

ngs RNA-Seq hisat indexing • 7.0k views
ADD COMMENT
0
Entering edit mode

If you point to the exact files you want to use for your fasta, splicesites, and exons, then perhaps some kind person will simply index them for you so you only have to download the files. Having said that, with 4GB RAM, I'm not sure you'll be able to load the resulting indices into memory.

ADD REPLY
0
Entering edit mode

Are you talking about outsourcing to someone who has high computing power?

ADD REPLY
0
Entering edit mode

That or just use a more powerful computer.

ADD REPLY
0
Entering edit mode

or buy aws instance

ADD REPLY
0
Entering edit mode

Note that you can also download pre-built index files: ftp://ftp.ccb.jhu.edu/pub/infphilo/hisat2/data

ADD REPLY
0
Entering edit mode

That was my second option. I was trying to do it by my own.

Clarification: I need to analyse RNAseq data for differential gene expression. As my plan was to build index with splice-sites and exon, which of the index will should I download? If I understand it must be genome_tran. Am I in the right direction?

ADD REPLY
2
Entering edit mode

From hisat issue #16, requirement is high (copy/pasted). But this is 2015 post. Probably indexing is improved:

HISAT2-build used about 160 GB of RAM when building indexes using the human reference genome and 12.3 million common SNPs,

From here, you can approximate the requirements. In the same thread post dated oct,2016, 64gb memory was not sufficient to index hg38 genome.

From the core of U of Pitt with HTC cluster, copy/pasted warning:

WARNING: In order to index the entire human genome, HISAT2 requires at least 160GB of RAM, and several hours.

ADD REPLY
0
Entering edit mode

The available Ensembl index is for release 84 dateing to Feb 2016; whereas current one is release 90. How far will that have an impact on my analysis?

ADD REPLY

Login before adding your answer.

Traffic: 2090 users visited in the last hour
Help About
FAQ
Access RSS
API
Stats

Use of this site constitutes acceptance of our User Agreement and Privacy Policy.

Powered by the version 2.3.6