Entering edit mode
6.5 years ago
hafiz.talhamalik
▴
350
I am working on a project where i have to do human genome assembly. I have around 100 samples. Computer with how much specification will be required for fast and good assembly. Anyone who can help me in this regard. Any specific model ?? Thanks is Anticipation
You should provide some additional information. What kind of data are you working with? How much data do you have for each sample? Are you planning to do de novo or reference based assembly?
Without specifics general recommendation would be as much RAM as you can get. 1TB (or more) would not be out of question assuming you would be assembling the samples sequentially.
I am working on Human DNA Sequencing Data, and i am doing de-novo Assembly. and i am planning to do assembly sequentially as well as parallel. I am very much clear about HDD, but i am not sure about minimum RAM and Processor requirements.
That is still not enough information. Is the data from Illumina?
Minia may be your best bet if you need something that can run on light hardware requirements.
Yes Its Illumina Data. and paired end data
i am willing to buy a heavy system no issue on that..
I have assembled human and rats with 256 GB RAM using PE data only (no LJD/ mate pair libraries).
This takes a while, at least 1-2 days per sample on ~32 cores. Haven't done it frequently though so can't give exact details.
Of course this all depends on the assembler - soap denovo works very well, and the stats here are based on that approach.
It might make sense to do alignment based approaches (BWA etc) with that amount of short read data. The assemblies will be very fragmented (like most fragments potentially 1-5 kbp, meaning fragmented genes). For alignment you would need less RAM, ie. 24-64 GB would be ok. More processors are always better.
Thanks for your kind help.