Forum:computer specs for open-source tools
2
0
Entering edit mode
7.3 years ago
bioguy24 ▴ 220
I am in the process of building a computer. My resources are limited but is an I5 with 16gb, 500gb sata hdd ok for tools like bowtie2, bedtools, picard, etc... I am planning to put 64-bit ubuntu 14.04.3 on it as well. The source of the data will be genomic resequencing of the medical exome (4500 genes) Any thoughts/suggestions. Thank you :).
next-gen Forum • 1.6k views
ADD COMMENT
1
Entering edit mode
7.3 years ago
DG 7.3k

How limited are your resources? I have a couple of suggestions from experience:

  1. A 500 GB SATA, unless you have a large NAS or other storage device you are routinely putting data on isn't nearly enough. Even if you do have a NAS It still is pretty small by the time you get various reference and annotation files downloaded, and start working with multiple samples. Assuming you're not just going to be working on a single exome. You really need to bump that up to a few TB If you don't have any other sort of place you are backing up key data too, you probably also want to consider some sort of RAID set up. Losing data sucks, and drives WILL fail. And usually at the worst possible time too. IMO Western Digital is probably the vendor of choice right now for drives. Unless speed of the disk is a major bottleneck WD Blacks are good for desktop drives and workstations. WD Re4 is an enterprise level drive that is pretty similar in price (4TB is about $249 US). As suggested a SSD for the Operating system, key resource, files, etc is also very nice. 120GB is about the sweet spot for price/performance right now.
  2. 16 GB of RAM will work... barely. I am currently using a workstation with 16 GB after coming from a machine with 48 GB. I have to remember because even though I have the same number of available threads as on the old machine I often have to run with less if I am passing 4 GB of RAM max to a thread in the GATK toolchain for instance. It will work but if you can get more, get more.
  3. You should really look at a workstation with the Xeon processors. The performance difference is huge, even with a similar number of cores and clockspeed, between a Xeon and i5.

What does your budget look like?

ADD COMMENT
0
Entering edit mode
I actually do have a xeon processor, with 64gb, and a 1tb hard drive. That unit has windows 7 on it and my original plan was to partition that with ubuntu 14.04.3 but that has been very difficult as ubuntu does recognize windows to run alongside. I have put in another 256mb hdd and will try installing both to that. I'm not sure why but other installs were easy (used windows xp and ubuntu 12.04). Anyway, I dont really have a budget, just trying to get an idea. Thank you :)
ADD REPLY
0
Entering edit mode

It is strange that you ran into problems dual-booting Ubuntu 12.04 and win7. I had that configuration in my laptop for years and it was easy to set-up, even for a beginner like I was at the time.

Regarding your original question, even though your original configuration, i5+16GB RAM+500GB HD might just work, you will soon:

  1. Run into disk space problems. Adding external HD might attenuate that, but afaik reading and writing from those might be slow.
  2. Grow frustrated at (i) the time it will take to perform some operations and (ii) the shortcuts to save time.

If you have the budget, follow Dan's advice and buy more power. For instance, if you decide to do some RNA-seq analysis and use STAR for that, it will require ~30GB of RAM. So think ahead :)

ADD REPLY
0
Entering edit mode

It should, that said my most recent personal laptop while I COULD re-adjust all of the partitions, there were a lot of small system partitions that Windows loves to put on the machine and make it a little confusing at times. But there is no reason why you can't dual-boot the two. That said, if you rarely need Windows, or only for some very specific tasks, just installing Ubuntu on the workstation with Windows inside a VM can be a much better option. Saves you needing to reboot (which is handy if you have a long running job on Linux but you just need to use something small in Windows for a bit). Performance can also be pretty good. I've been doing this for years just to run GenomeStudio for some genotype data file processing for instance or currently because I am now at a hospital to have access to the managed desktop with shared drives, etc.

Something to consider.

ADD REPLY
0
Entering edit mode
7.3 years ago
kloetzl ★ 1.1k

Get a SSD (100-200 GB). A hardrive is legacy hardware for storing huge data files like genomic data an backups. For day-to-day usage a SSD is really worth the money.

ADD COMMENT

Login before adding your answer.

Traffic: 1923 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