Commercial Software vs. Freeware

[dnart]

Our lab is trying to decide whether to go with commercial software, freeware, or a combination for our NGS data analysis.
I am not a programmer or bioinformaticist, and I'm a bit overwhelmed by the number of tools out there and the whole command line / Linux environment. I've tried Galaxy online as I thought I don't need to know command line or Linux to use it, but again I am a bit overwhelmed by all the tools and options included in it that I have no idea how to configure.

So my question to the community, is which is more cost effective? Hire a bioinformaticist to help me install, configure, and run these tools, or go with a commercial software package? Why do you feel this way? Note: I am not looking for recommendations on a particular package, just a general discussion.

[turnersd]

Sounds like you've discovered the true cost of "free" software. I'd say I use free software for 90% of my work, but there are advantages to paying for something, namely, the support. I wrote a blog post discussing free vs paid software after I evaluated Golden Helix SVS (I don't use the software any more, as I rarely do genetic epidemiology / GWAS any more).

http://gettinggeneticsdone.blogspot.com/2012/05/stepping-outside-my-open-source-comfort.html

[GenoMax]

You may have already looked at a previous thread with discussion on this topic: http://seqanswers.com/forums/showthread.php?t=23905

A lot may depend on your local "environment". It would be hard to enumerate that setup in a discussion forum so ultimately you will need to decide what makes sense in your case. I would definitely suggest talking with local IT resources to gauge their support, specially with respect to infrastructure (networking, storage, clusters).

Are you referring to "hiring a bioinformaticist" as a one time consulting gig or will there be enough work to keep the person employed/interested long term. Something to think about.

[dnart]

My local environment is that we all have either PC's or Macs and none of us are familiar with Linux or scripting - we are medical researchers not bioinformaticists. To get substantial IT resources would require paying for someone to provide that. So, yes, by "hiring a bioinformaticist" I am referring to a long term gig, which I guess is pretty clearly much more expensive than buying a commercial software package that comes with technical support. To the previous poster, yes, we are concerned about the *real cost* of "free" software.

[GenoMax]

My local environment is that we all have either PC's or Macs and none of us are familiar with Linux or scripting - we are medical researchers not bioinformaticists. To get substantial IT resources would require paying for someone to provide that. So, yes, by "hiring a bioinformaticist" I am referring to a long term gig, which I guess is pretty clearly much more expensive than buying a commercial software package that comes with technical support. To the previous poster, yes, we are concerned about the *real cost* of "free" software.

dnart: If you are planning to do a substantial number of samples (and their analysis) then you should look at investing time/effort into setting up centralized infrastructure. Using standalone PC's/Mac will work to some extent but you will soon find yourself looking for storage/compute cycles. It may seem attractive to buy a bunch of externals disks and use them as your long/short term storage but the downsides may bite you at the worst possible time (e.g. having to recover some irreplaceable data and finding problems with the disks).

Hiring a bioinformatics person is just one part of the equation and may be the first thing you will do. But please budget for/plan to put some reasonable IT infrastructure in place if you expect the informatics person to deliver results.

By "environment" I was mainly referring to administrative/accounting chain of command at your place of work and their tolerance for a substantial amount of expense that is bound to be incurred as you move forward. Think of it (or sell it) as a part of the total package (buying/paying for the actual sequencers or a sequence provider if you get your sequencing done elsewhere).

[FractalExpression]

This is an issue my clients and I have dealt with in both academic & industrial settings. The "cost of free" software is high in-terms of usability, while the current commercial packages lack flexibility (not to mention the learning curves associated). What you want & need is become comfortable with terminal, as a modern researcher this is fast becoming a necessity.

With this in mind some friends and I have been putting together what you could call "Unix training wheels"...essentially like an old video game like zelda that can walk you through performing complex bioinformatics tasks by simple Q&A dialogue.

Check out the screenshot,

If this looks like something you'd be interested in, just say so and I can send you the .app and help you get started.

And if anyone else wants to give us some feedback, I'm all ears

[dnart]

As a biologist, I'm not really interested in learning Unix. As far as buying a bunch of infrastructure and learning to use terminal, I think a more attractive alternative for me might be to go with a commercial solution where I stream my data to their hosted solution or buy my own resources on a cloud such as Amazon. Does anyone have any experience with this route?

[gsgs]

I never learned UNIX/Linux, but I do feel that command-line programs
(cmd.exe) are just better than mouse-clicking stuff.
More flexible, can run from batch, easier to adapt to new tasks,
easier to correct/repair/expand. Easier to describe to others,
how to use - just print the command in email -
no list of instructions what windows may open, where to click.
And I do remember, that I felt sad )and still do) when Windows appeared
to replace DOS-command-line utilities

[xied75]

I never learned UNIX/Linux, but I do feel that command-line programs
(cmd.exe) are just better than mouse-clicking stuff.
More flexible, can run from batch, easier to adapt to new tasks,
easier to correct/repair/expand. Easier to describe to others,
how to use - just print the command in email -
no list of instructions what windows may open, where to click.
And I do remember, that I felt sad )and still do) when Windows appeared
to replace DOS-command-line utilities

No need to be sad, the new CMD is PowerShell, so powerful thus the name.

[gsgs]

http://en.wikipedia.org/wiki/Windows_PowerShell

sounds complicated with lots of unknown TLAs
will it run my old programs ? I couldn't find it yet

[chadn737]

Large scale data is not going away and using commercial software means that you will always be playing catch up. Once you learn how to manipulate files in Unix, you'll realize how much time you've been wasting over the years trying to do this stuff in excel or by hand.

It behooves any Biologist to get with the program and start learning how to do things in a command line. Its another skill in your toolbox, more useful than most wet lab techniques and increasing skill in demand.

[dnart]

I don't think you are getting it. I am not going to learn command line programming, instead I am going to focus on learning more biology. The question was whether it is more cost effective to hire a bioinformaticist (~$75k/year) or buy commercial software (~5k/year).

[chadn737]

I don't think you are getting it. I am not going to learn command line programming, instead I am going to focus on learning more biology. The question was whether it is more cost effective to hire a bioinformaticist (~$75k/year) or buy commercial software (~5k/year).

I do get it. Its not a tradeoff of learn biology or learn programming. Learning the programs and the statistics can make better biologists. Its not an either/or dilemma.

But, if its Bioinformatician vs software....part of the problem of simply buying point and click software and not learning anything about the programming is that its easy to stick data through the software and not know whats going on or how that affects your end results. It's like a bench scientist only using kits with no comprehension of what is going on. Yeah, you get a result, but you don't know if its the right result.

[HESmith]

I don't think you are getting it. I am not going to learn command line programming, instead I am going to focus on learning more biology. The question was whether it is more cost effective to hire a bioinformaticist (~$75k/year) or buy commercial software (~5k/year).

If you frame the question like this, then the answer is obvious.

But perhaps a better question would be "What is the most cost-effective way to extract biologically meaningful information from next-generation sequencing data?" Given that you find the number of options in Galaxy daunting, and you have no interest in learning command line programming, then hiring a bioinformatician seems essential. For example, an often overlooked aspect is quality assessment of the data prior to analysis, but most commercial software packages lack tools for this application. Furthermore, without the expertise of a bioinformatician, how would you evaluate the relative merits of different software (commercial or free), or understand how different parameters in your analysis pipeline might affect your output? A good informatician can also help in the experimental design (e.g., knowing how many replicates you need to obtain statistically significant results). Finally, data analysis takes a lot of time (much longer than performing the benchwork and generating the sequence data), so you should consider how much of your own time you're willing/able to commit to the effort - time that would not be available for learning more biology.

My two cents' worth, from a bench scientist doing his own analysis.

[dnart]

We do have a statistician already working with us, but he is not an expert at linux and command line tools for alignment etc. In the past I had no trouble analyzing my microarray data using Genespring at first, and later switching to Partek's Genomics Suite. I am quite familiar with doing the normalizations, t-tests/ANOVAs, clustering, gene ontology / pathway enrichments, etc. So I don't think it is a matter of not being able to analyze my data, it is more a matter of getting my unaligned data through the aligners and quantified at a level that I feel qualified to do the analysis myself.