Hello Hasnahana, you gave some information about your object of interest, although there is a lot of factors that should be considered while evaluating required computational power for this task. Do you mean de novo assembly or you have some close related mammal that you can use it's genome as reference? Also do you have some preferences using software? Different tools require different computational power. What NGS technology do you use and how long reads are?

Hello. The technology used can be Illumina and we have to do a de novo assembly. TopHat and Cufflinks are the most commonly used, so suppose I use them; then what are RAM and CPU core requirements?

Just to prevent further wrong assumption:

TopHat and Cufflinks are used to analyse RNA-Seq experiments, they have rather weak requirements regarding RAM and CPU.

You might rather be looking for tools like:

VELVET,SPAdes, ABySS,etc.

Read here for a short introduction:

https://en.wikibooks.org/wiki/Next_Generation_Sequencing_%28NGS%29/De_novo_assembly

I must admit that I do not know about performance requirements but you might need a lot of RAM.

According to this (http://listserver.ebi.ac.uk/pipermail/velvet-users/2009-June/000359.html) you could calculate RAM usage depending on the size of the K-mers the amount of reads and the amount of genome and the size of your  reads.

Best,

Peter

Sorry that was a mistake on my part. Yes, I will use velvet  and ABySS.

We just did a de novo assembly of C. elegans (100Mb), and if I recall we used ~100GB RAM and 10 or 20 cores. It took a few hours or a day. For a larger genome like yours you'd probably want 100-200GB RAM, but CPU cores don't matter nearly as much. I'm sure you could get away with 4-8.

David, do you have any idea if de novo transcriptome require as much or more specs?

The RAM requirement for transcriptome assembly is typically much lower. I've run transcript assemblies on my desktop computer (32Gb RAM, 8 core) with tens of millions of paired-reads in less than a day. Note: The time required depends quite a bit on the software and how complex the transcriptome is.

Did you sequence DNA or RNA, or both?

That sounds promising. But was your assembly de novo? 

In this paper you can find informations about RAM usage, time of analysis etc. during de novo transcriptome assembly for different tools. Take a look.

http://bmcbioinformatics.biomedcentral.com/articles/10.1186/1471-2105-12-S14-S2