Help a graduate student going into Bioinformatics looking for a new personal laptop. Should I get a Mac or a PC?

Question

[Angelo Victoria]

Hello, Biostars!

I'm new here and I'm really interested in pursuing bioinformatics. I'm currently a first-year MSc student working in a project involving genomics. The actual computer work will be concentrated next year so I have some amount of time to learn everything I can manage. My undergrad is a classic, broad BSc in Biology so I have practically zero experience doing bioinformatics, save for a molecular phylogenetics exercise we did in my elementary genetics course.

So my question is, how different is OSX to Windows in the context of bioinformatics? I'll be buying a new laptop to replace my old one and I'm thinking of moving to OSX (MacBook Pro 13-inch 2015, 2.7GHz Core i5, 8GB RAM) instead of using Linux over Windows (that's the virtual machine, right?). We have a core facility that has supercomputers to run analyses in so I'll be using the laptop as a personal laptop for coursework as well as to write code when the time comes and access the supercomputers to run my data in them. I've read anecdotes that Windows is pretty useless in bioinformatics. Can you shed light on why? And are Macs enough, even though they are less powerful than PCs?

Please treat me as a total newbie and feel free to correct me on any mistakes in my line of thought above. Thanks!

P.S. apologies in advance the unnecessary use of tags. I'm still learning the etiquette of this site. :)

[Biomonika (Noolean)]

Definitely get a Mac computer

Answer 1

[John]

When I was 17, I decided I wanted to be an illustrator.

I spent hours researching what tools I'd need to reach my goal. I settled on three must-haves - a Prismacolour Premier 0.05, a Sakura Pigma 0.3, and an array of Staedtler Pigment liner pens for filling in. Then I'd need a drafting pen, a bunch of pencils from some hipster pencil company I've forgotten the name of (but their graphite, very importantly, came from Morocco, while the wood was some kind of fine-smelling cedar... ), and finally a whole bedroom of craft papers in different shades of off-white, grades of coarseness, softness and stiffness.

After spending about £200 on pens and pencils, I then discovered something quite interesting about myself - I hate drawing. Like, I find it really, really dull. Also, I suck at it. These things are probably somehow related to the fact that I spent more time buying pens than actually drawing. So I guess what I'm trying to say is....

[h.mon]

+1 for the drawing - you do have some talent.

[Charles Warden]

While this answer did make me chuckle, I feel obligated to bring attention to the fact that the response with the most +1s doesn't actually address the original question.

[lh3]

Voting reflects the most popular answer, but not the most correct or relevant answer.

[John]

In defense of my terrible drawing, the most correct answer is not always the most useful either - particularly if the question is an XY problem to begin with :)

[lh3]

Right. Voting reflects the most popular answer, but not the most correct or useful or relevant answer.

Answer 2

[Devon Ryan]

Do not use Windows.

Either get a Mac or get a PC and install Linux. Attempting to use Windows for anything serious in bioinformatics is a recipe for stress.

Edit: A Mac is fine. It's sometimes slightly more difficult to install something, but only slightly. Usage will be the same as under Linux. I actually used to have a Mac as a laptop previously and had no issues doing anything on it (using R, programming, writing papers, etc.). The biggest benefit of OS X over Linux is that you can use Office, so writing papers with colleagues is a bit simpler (that's what they're likely to be using).

Edit2: I should note that one certainly could use Windows to access remote clusters where the actual data processing will be done. However, you'll likely find that it's convenient to not always be tethered to these remote systems and instead be able to just do simple coding and such on your local machine. That's all vastly simpler in Linux/OS X than Windows. The only benefit to Windows is the MS Office package (which is an unfortunate requirement in much of academia), which is also available on a Mac.

[Ram]

Adding to Devon's answer, use Parallels on a Mac to run a Windows VM for office (if you're particular about that - Office in Windows can be vastly superior to that on Mac).

Windows is indeed almost completely useless in bioinformatics. Most PCs use HDDs, Mac has more prevalent SSD options, and that could make file I/O intensive bioinformatics tasks easier on a Mac. But, there's the difference in price tag.

The best part is, it is super simple to get GNU binaries as the default -nix binaries on a Mac. This is nearly impossible on PCs. GnuWin32 gets you close, but offers nowhere near the vast features as OS X does.

Even if you get a PC, a VM is not as good as dual boot. Get a Mac, it'll make your life easier.

[andrew.j.skelton73]

The MS Office Beta for OS X isn't too bad, https://products.office.com/en-us/mac/mac-preview - Definitely bridged a much needed gap

[Keith Callenberg]

You can use MS Office on Linux, just not natively. I made it through grad school running MS Office 2010 via Wine and had no issues. https://appdb.winehq.org/objectManager.php?sClass=application&iId=31

[Daniel]

Wine very much doesn't work with Office 2013 though. I needed to set up a VM to do it.

[Keith Callenberg]

Yeah I included the AppDB link above to address exactly that point. I guess you can't get academic licenses for 2010 anymore.

[zizigolu]

yes all right

Answer 3

[Alex Reynolds]

I realize this won't be a popular view, but, while more expensive, Apple simply makes the best hardware for the dollar (or other currency unit).

For laptops, you cannot do much better than a Retina MacBook Pro or a non-Retina MacBook Air. The Retina MacBook would be a good choice when the second- or third revision comes out.

There are many "sub-US$500" (or local currency equivalent) Linux options, but they do not compare in build quality, performance and support with Apple hardware. If taken care of, MacBooks can last 5-6 years, while you might expect to have to replace a sub-$500 laptop within 2-3 years.

Further, all the non-Apple laptop alternatives that have added the same hardware features as MacBooks tend to be made with lower quality components and - weirdly enough - cost more than Apple laptops.

A laptop running Windows is not the right choice for bioinformatics, which leaves OS X and Linux. I am a strong proponent (and author) of free open source software, but for doing productivity work, OS X simply offers many more and - more importantly - much higher quality productivity and illustration software options than Linux: Keynote, OmniFocus, OmniGraffle, Papers, Adobe Illustrator, etc.

As OS X is a BSD Unix, virtually all open source tools available to Linux users are available to OS X users with a recompile or package install with Homebrew or MacPorts. A lot of scripts can be run as-is, because Python, Perl and so on are built into the operating system already.

It can also be easier to connect wirelessly to some projectors with an OS X laptop, for lack of client software support on Linux. So giving presentations can be less of a disaster.

Linux is excellent for data centers, computational clusters, and other server applications - no question - but it is not as good a platform for the laptop user as an OS X-based Apple laptop. Unless you need to carry the data center in your backpack, Apple makes the best tools for portable computing.

[Damian Kao]

I agree. I think most people tend to think hardware in terms of raw performance power and not build quality/portability which is just as important. It's unfortunate that Apple is such a polarizing company. People either love it or love to hate it.

[Antonio R. Franco]

I don't agree.

It is true that many bioinformatic programs can be compiled for the Macintosh, as its operative system is actually UNIX. But there are many other programs that lack the code that will allow a Macintosh compilation. Thus, you eventually will be working with a Linux system embedded into your Mac with the same processor and memory you can have after buying a cheaper Windows based computer

Is also true that you can run SOME bioinformatic applications in your local computer when the hardware needs are not high, but most of NGS applications cannot be run on a local computer: you need to work in the cloud. Working in the cloud is the clue. And this can be accomplished with a cheaper PC

[alec_djinn]

Please make some examples of programs that you cannot compile on Mac...

[lh3]

Docker ;-)

[Alex Reynolds]

The screenshots on the Docker project page appear to have been taken under OS X (Yosemite): http://docs.docker.com/project/set-up-prereqs/#fork-and-clone-the-docker-code

[Devon Ryan]

Apparently on a mac you need to run a linux virtual machine to use docker. This wouldn't exactly be a deal breaker for using OS X for me personally, but I'm not exactly a huge docker fan (though we use it internally for some things).

[Alex Reynolds]

I use the version of Docker that runs natively on Mac. Seems to run okay.

[Devon Ryan]

Cool, looks like they need to update their documentation then :)

[lh3]

Docker client is a mac binary, but it has to talk to a docker server which is linux only. The way docker works on mac is that it sets up a small linux VM via virtualbox/vmware, launches the docker daemon in the VM and makes a port open for communication. The docker client on mac talks to the docker daemon via the port. All the heavy lifting is done inside the VM.

[Alex Reynolds]

Sorry, my misunderstanding. It sounded like it was said that you have to run VirtualBox or similar in which you run the Docker client, which doesn't appear to be the case. I guess it is an issue of semantics.

Answer 4

[User 59]

I don't think it matters personally, I've switched between PC and Mac in various work environments (most recently 4 years Windows only, mandated by the company!). SSH and Virtualbox will get you wherever you need to go! All your heavy lifting will be done on a server/in the cloud anyway.

Where I work now, it's a pretty bioinformatics heavy institute, and I can assure you not every 'serious' bioinformatician is toting a Mac.

[Ram]

If the institute you're studying at or working for has a cluster, the personal machine seldom makes a difference. However, OP was talking about grad school, and some grad schools might not have clusters. Also, not sure if OP is willing to pay for Amazon's flexible compute services. In this case, a Mac+homebrew will make quite the difference. If geeky enough, OP will feel no difference between a Mac+homebrew and PC+Linux.

Answer 5

[SES]

I will go against the grain and say do not buy a Mac*. Whatever convenience you gain is not worth paying 10X the price, especially with cloud computing so readily available now (and cheap). It can be hard to undo a broken installation on a Mac but with cloud computing you just click delete. Also, a Mac is not so user-friendly out of the box as people would have you believe and a beginner will have plenty of difficulties trying to compile programs and install packages. You will have any easier time with bioinformatics software on a Mac than a peecee for sure, but the Mac system has plenty of vagaries that are unique and challenging.

My main opposition to the Mac idea is that I believe in open source and I don't think we should have to pay so much for a system, then pay to update the OS in six months, pay for the office software, etc. You will do your actual computing on another machine anyway, so in my opinion it is not so critical. That is, unless you are concerned about looking cool, then definitely get the Mac and the apple watch and skip a few meals. I'm kidding of course, the best solution to being productive is probably to stick what you are familiar with for everyday tasks and use cloud computing and/or a VM for work.

*I'm writing this on a Mac :)

[Alex Reynolds]

Apple doesn't charge for OS upgrades or for their productivity software. I don't think that they have done so in a while.

[SES]

At first I didn't believe you until I looked this up. If you owned Mac computers for many years you know that iWork has never been free and the same goes for the upgrades. The pricing history on both the OS upgrades and iWork is on Wikipedia, which says if you purchased a computer in October '13 or later you are eligible for a free download of iWork. That is nice that you don't have to pay for these things now, and it is something that would make me consider getting another Mac.

Answer 6

[Daniel]

Very subjective, but I think the universal truth is that with a PC/'non-mac' you get more hardware for your money.

The other universal truth is that it's much harder (but not necessarily impossible) to do any medium-to-serious bioinformatics in a windows environment, compared to a linux/mac.

Personally, I find linux much more applicable and easy to use (particularly BioLinux), and in my limited experience of Macs I found myself getting frustrated with software versions not being compatible with particular OSX versions.

Answer 7

[Damian Kao]

I think the general consensus is that it doesn't matter what hardware you get since most of your heavy work will be done on a remote server. The only thing that really matters is to get some flavor of linux for your OS.

Answer 8

[JC]

My 2 cents:

• If you only want to start doing bioinformatics (simple coding, web browsing, simple databases) doesn't matter what you use.
• If you are starting to code and test more complex tasks, you should be using Mac o Linux, Windows is not an option.
• If you have a large data set to read, parse, analyze and do-stuff, your hardware and OS will become really important so Linux is your option (I use Fedora BTW), except when you have to do all the analyzes in a server, cluster or cloud in that case you can use Mac, Linux, Windows or even a tablet.

Answer 9

[Antonio R. Franco]

I will go for a PC with a Debian Linux such as Ubuntu, but the PC at least, have to have a minimum of RAM installed (16Gb or even more) and an high class I7 processor.

You will pay a fraction of what you have to pay for a Max, and you can use local programs and also cloud computing for more serious tasks

Look into omictools.com. Then you can get an idea of how many bioinformatic utilities are coded for Linux (the vast majority of them, indeed)

I don't like Virtual machines run on Windows or Macintosh. They are a operative system working under another operative system and this means limitations

Answer 10

[lh3]

A few more things to consider:

1. If you buy a mac, you can relatively easily install linux and windows on it along with OSX. If you buy a PC laptop, it is nontrivial to install OSX.
2. If you often write python/c code to work on linux, you can easily develop and test your programs offline on mac/linux. This is important to me. On windows, such offline development is doable, but will be more clumsy - windows shell is weird; cygwin is inconvenient; vm drains battery - my personal opinion.
3. I reiterate Devon's point: you will need MS office if you often work with biologists. They will give you drafts in MS Words which are best edited with Words.
4. If you read the "use-this" series or go to any bioinfo conference, you will find the majority use macbook pro/air. Nonetheless, choice of laptops is a very personal thing. A laptop for the majority is not necessarily the best for you.

Answer 11

[Biojl]

I disagree with a lot of people here regarding hardware. Hardware is also important in a laptop, try doing some data analysis / plotting with the data processed by your high-end cluster/cloud in R with a low-end laptop (limited memory and cores). At least in genomics that's not trivial anymore and it's cumbersome to do it in a cluster.

I would definitely go for a PC, better hardware for less money and then install linux (ubuntu or fedora). You can still install OSX as dual boot if you want to be fancy.

Answer 12

[h.mon]

Browse and read the series "Uses this" here. One lesson you will learn is anything may be used for "serious" bioinformatics, several people develop different workflows.

edit: and very often, one will use a variety of hardware to acomplish a task.

edit2: my personal setup: a cheapo laptop and a beefier one, Debian on both, data synchronized between them. If you decide for Linux, do some research for hardware support (drivers and / or malfunctioning hardware) and easy of installation (secure boot issues). For docs / spreadsheets I use LibreOffice, but had problems with format screw-ups when sharing files a few times. I keep Windows mostly for internet banking, as I could not install the bank secure software - so yes, usability in the "real world" is sometimes a problem for linux.

[h.mon]

In addition, not related to your question but useful as you want to pursue bioinformatics, this rant talks about which programming language you should learn (maybe yourn next question). Be warned, this blog has a very polemical post, one which is not very encouraging for someone starting on the field of bioinformatics - it should not discourage you but instead help you avoid some bioinformaticians weaknessess.

[zizigolu]

till now i did all of my analysis by windows but i faced with too much difficulties..

[Ram]

If your data is local (in the same machine you are running the analysis on), Windows is the worst choice you could have made.

Answer 13

[Charles Plessy]

For the Linux distribution, I recommend Debian, because if you just install Debian Med's packages like med-bio or med-cloud, you will already have a large number of popular bioinformatics tools installed without effort.

[Ram]

Umm, Bio-Linux?

[Daniel]

http://environmentalomics.org/bio-linux/

[Ram]

Yes, I was suggesting that Bio-Linux is a good alternative to Debian/Ubuntu.

[Daniel]

I was just adding a url for quick access! :-)

Answer 14

[mikhail.shugay]

As others suggested, the choice of hardware is not really important as most computation-heavy work should be done on UNIX servers. So in case of a Macbook, I would suggest getting Macbook Air not Pro as it is really light and could be easily used in trips, carried over the campus, etc. The benefit of Mac (compared to other ultrabooks) that complements its usage in the field conditions is its great touchpad.

Answer 15

[alec_djinn]

Definitely a Mac, with 12 or 16 Gb of RAM. It is not only matter of hardware. You will need to run scripts, write code, read webpages, edit graphs, work on different softwares in the same time. What is nice on a Mac is that you can turn off your computer and expect to have everything back exactly how it was before powering off. You can move across desktops smoothly while everything still running in the background, you can open an excel file of 50MB and search on it without problems. Windows just freezes most of the time you try to do so... Stability of the system is unbelievably important when you are working for days with different apps. I tried many times to run the same memory crunching scripts on OSX an Windows, with machine having similar CPU and the same RAM. Windows simply freeze! OSX goes on like bull. Believe me, you do not want to see your script crashing after one hour of calculations because someone is calling you on Skype in the meantime sucking the last few bites of memory. Also it is not matter of cost, to spend few box more for a computer that makes your life easier is a great deal. I saw countless people (me included) willing to save money buying a plastic calculator and ending up after 6 months with poor battery life, 2 antivirus running in the same time, and just very slow, almost inoperable, windows system. If you really go for bioinformatics you will spend most of your day time with that computer.. choose a damn good one!

Answer 16

[poisonAlien]

I had same dilemma couple of months ago. First I thought I would buy any laptop and install Linux on it. But I have had terrible experience with my Dell xps system running on Linux as it gets overheated easily. This is also a known issue and many have posted this on ubuntu forum (google: laptop gets hot and shuts down ubuntu site:ubuntuforums.org). After doing some research and going thorough uses-this thread, I bought macbook pro 13' retina, and I am happy with this for lot of reason. It has Xcode, homebrew, most of the core bioinformatics tools can be compiled on mac. Plus I can use office utilities, endnote, adobe like in windows and it has battery life of ~10 hours.

Answer 17

[deepue]

Hi,

I am a Bioinformatics Grad student with very basic knowledge. I was in the same confusion before purchasing the laptop few months ago. As most of them mentioned here, you can't run heavy computations(e.g., NGS data) on your local computer/laptop. As there will be servers available in most of the universities for Bioinformatics computational work, it doesn't matter. Almost all the Bioinfo related software is compatible with every platform, so go with the one you are comfortable with. After working in the field, you will understand which one is better. With these assumptions, I have purchased laptop with Windows 8 and working on servers for Bioinformatics work through windows. I might be wrong in my assumptions but I didn't face any major obstacles due to this.

Thanks

[zizigolu]

yes, I think so. I am also practicing with windows 7 but for serious computation I need to use the my institute systems

Answer 18

[abrahamdsl]

Hi there fellow UP student! I'm from Los Baños campus and I'm now working here beside the university as a Bioinformatics assistant even though I graduated BS CompSci and not having an inkling much on Bioinformatics before getting the job.

So far, I would recommend a PC where you can install and use Linux preferably Ubuntu as the main operating system. With our government suddenly having a focus on this field, I am hearing that a few high-performance computing clusters are set up here and there every now and then, so you'll tend to do your workload on these machines accessed using PuTTY rather than on your workstation.

[Angelo Victoria]

Oh hi! I guess I'm from the other side of the spectrum as I have virtually zero knowledge in computer science. I'm from Diliman but we occasionally travel to the Institute of Plant Breeding for our project. My current introductory course in Bioinformatics allows the use of Windows as of the moment but I know that more advanced workflows would require Linux so I opted to install Ubuntu alongside Windows (can't live without Office to write papers!) until I get a new laptop.

[Biomonika (Noolean)]

There is Office for Mac that works just as good as Windows version:)

[Ram]

Impossible. There were features in Office 2010 that Office for Mac 2011 was incapable of doing. The new version would need to have improved by leaps and bounds to catch up to 2010, let alone the silky smooth 2013.

[abrahamdsl]

Yes, for papers. Microsoft Office is essential.All of us have requested from the IT department that our laptops be dual-OSes/dual boot.