How to set up a computer for bioinformatics? Any hints..

leonardo_ca · 12-16-2017, 11:15 AM

Hey there,
I know this question has been around a few times, however the threads I could find are relatively not so updated and as Bioinformatics and computer evolves quite fast, I decided to ask it.

So, I am up to start working with Bioinformatics for a lab and that will be a new "line of research" there, which means we are buying computational resources too. We will do bacterial sequencing using the Miseq platform. There will be 16S amplicon sequencing for bacteria identification (multiplexing at maximum) AND sequencing for genome assemblies to create a database of the sequenced species. I am considering demultiplexing and running Qiime for species identification; and for assemblies running AbySS and quality checking.
I am experienced with Bioinfo (plant RNAseq) but never had to buy a computer and that`s my problem now. Could you help me with some hints on what to buy at a medium budget? I believe should go for something with lots of RAM and storage...

I though in two main lines:
1) a powerfull but not that expensive machine:
Dell Precision T3620
Xeon E3-1225 v6 (4 cores, 3,3 GHz, HD Graphics P630)
32 Gb UDIMM DDR4 ECC 2400 MHz
NVIDIA QUADRO K620 - 2GB
1 TB Sata Storage + cloud storage (I am considering cloud storage so 1TB is for instant storing)
Ubuntu 16.04

2) cloud computing
Google or Amazon cloud computing on a similar than "1" image. It's not that cheap but, an option. Still, never worked with cloud computing so do you think that is reliable for bioinformatics? Do the assemblers, mappers etc run on the cloud?

Well, I hope you guys could help me on choosing a reliable configuration.

Cheers and thanks,

GenoMax · 12-16-2017, 11:58 AM

Devil is always in the details of how much sequencing you would be doing. Reagent costs for sequencing add up rapidly compared to cost of the hardware you are going to buy for computing.

I am not sure how much the Quadro card you have listed is but I would let that go and get more RAM if you can. RAM is one thing that you can't find an alternative for when you need it. While cloud computing sounds fantastic you should stay away from it if you work in a small individual lab.

leonardo_ca · 12-16-2017, 12:33 PM

Dear GenoMax, thanks for replying.
Good point, we can make our way for get rid of low processor capacity and storing, but for RAM the only solution is buying more and more.
The lab is not exactly small, thing is that I am the only one who will doing bioinformatics there, because they used to hire such services. They were doing R&D on bacteria species discovering and cultivation but sequencing was out of their work and now sequencing and bioinfo will be done in the lab.

GenoMax · 12-16-2017, 04:35 PM

I would suggest that you get hardware that makes sense (don't forget a data backup solution, it can be as simple as multiple external drives). When in-house hardware is not sufficient for a specific project then you can turn to the cloud.

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Optimal bioinformatics computer configuration (budget 2500$)	AcoBetalow	General	3	01-08-2015 03:57 AM
MS in Bioinformatics or MS in Computer Science with a concetration in Bioinformatics?	Kimist99	General	5	11-26-2014 01:32 PM
Computer scientist with experience in bioinformatics	aurelielaugraud	Academic/Non-Profit Jobs	1	05-30-2012 04:59 AM
Computer Scientist diving into bioinformatics... where to start?	Fixee	General	10	06-23-2011 02:00 PM
Bioinformatics Computer: List your specs	peromhc	Bioinformatics	14	10-09-2009 11:21 AM

12-16-2017, 11:15 AM	#1
leonardo_ca Junior Member Location: Brasil Join Date: Dec 2017 Posts: 5	How to set up a computer for bioinformatics? Any hints.. Hey there, I know this question has been around a few times, however the threads I could find are relatively not so updated and as Bioinformatics and computer evolves quite fast, I decided to ask it. So, I am up to start working with Bioinformatics for a lab and that will be a new "line of research" there, which means we are buying computational resources too. We will do bacterial sequencing using the Miseq platform. There will be 16S amplicon sequencing for bacteria identification (multiplexing at maximum) AND sequencing for genome assemblies to create a database of the sequenced species. I am considering demultiplexing and running Qiime for species identification; and for assemblies running AbySS and quality checking. I am experienced with Bioinfo (plant RNAseq) but never had to buy a computer and that`s my problem now. Could you help me with some hints on what to buy at a medium budget? I believe should go for something with lots of RAM and storage... I though in two main lines: 1) a powerfull but not that expensive machine: Dell Precision T3620 Xeon E3-1225 v6 (4 cores, 3,3 GHz, HD Graphics P630) 32 Gb UDIMM DDR4 ECC 2400 MHz NVIDIA QUADRO K620 - 2GB 1 TB Sata Storage + cloud storage (I am considering cloud storage so 1TB is for instant storing) Ubuntu 16.04 2) cloud computing Google or Amazon cloud computing on a similar than "1" image. It's not that cheap but, an option. Still, never worked with cloud computing so do you think that is reliable for bioinformatics? Do the assemblers, mappers etc run on the cloud? Well, I hope you guys could help me on choosing a reliable configuration. Cheers and thanks,

12-16-2017, 11:58 AM	#2
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,088	Devil is always in the details of how much sequencing you would be doing. Reagent costs for sequencing add up rapidly compared to cost of the hardware you are going to buy for computing. I am not sure how much the Quadro card you have listed is but I would let that go and get more RAM if you can. RAM is one thing that you can't find an alternative for when you need it. While cloud computing sounds fantastic you should stay away from it if you work in a small individual lab.

12-16-2017, 12:33 PM	#3
leonardo_ca Junior Member Location: Brasil Join Date: Dec 2017 Posts: 5	Dear GenoMax, thanks for replying. Good point, we can make our way for get rid of low processor capacity and storing, but for RAM the only solution is buying more and more. The lab is not exactly small, thing is that I am the only one who will doing bioinformatics there, because they used to hire such services. They were doing R&D on bacteria species discovering and cultivation but sequencing was out of their work and now sequencing and bioinfo will be done in the lab.

12-16-2017, 04:35 PM	#4
GenoMax Senior Member Location: East Coast USA Join Date: Feb 2008 Posts: 7,088	I would suggest that you get hardware that makes sense (don't forget a data backup solution, it can be as simple as multiple external drives). When in-house hardware is not sufficient for a specific project then you can turn to the cloud.