what is the minimum system requirement for assembling long reads obtained from pacbio or oxford nanopore ?
I would say that the answer will depend on more what you are assembling (complexity of size or population)
For a 2 gb genome assembly?
How much data do you have?
We haven't sequenced yet. the organism have approximately 2 gb genome size.
In general, human genome can easily require a terabyte of RAM (or more) for de novo assembly so that should give you some idea. Having plenty of RAM would be a key. If your genome has repeats, ploidy (other then standard haploid/diploid) that would add to the complexity of this task.
Doing a de novo assembly with only long reads (which are going to have a high error rate 10-15%) is probably not recommended for a genome of this size. You would want to throw in some short read coverage as well.
A cluster computing facility is not mandatory right?
Cluster computing facility is actually no help here. A typical cluster architecture is designed to provide lots of CPU nodes but cannot give you more memory than what a given node already has.
nagendranp1991 : This paper would be of interest: https://www.nature.com/articles/nbt.4060
Check the compute section for details on what they used.