Graphical softwares!

[MoBi]

Hi,

I'm very new to the world of bioinformatics!
One question that I have to ask, what is the difference between the bioinformatics softwares that you can use in a windows-like fashion and those that you have to know a bunch of commands on a black screen to run them?! is it the speed of processing?

Also, do you know any good softwares that are very user-friendly to do reference assembly and mapping? (Is CLC bio one of them?)

Many thanks!

[westerman]

I would say that the major difference -- if indeed there is one -- is that command-line oriented software is often easier to write and easier to make run on multiple computers. Thus the people creating the leading edge software, especially the free public-domain software, will write command-line based software.

If you read the forum you will find that what is 'good' software varies from person-to-person and from job-to-job. Obviously if you can not or are uncomfortable with command-line based software then you will need to run windows-like software even if such software may be not as up-to-date as the other software.

As for CLC-bio, they do have trial software that you can download. Give them a try and see if they are suitable for you.

[dcjamison]

I think the major barrier to writing a GUI is the cost/benefit ratio. Java makes it simple to write a GUI, but what does that GUI actually do for you?

For example, think about what a GUI would contribute to a short read alignment program: Setting up your analysis would be clearly be easier, since you no longer would have to remember the at-best cryptic command line parameters and poorly described default behaviors we all put into our software. But after that, the GUI contributes little. During execution, the program outputs reassuring status messages (or less-reassuring error messages), but writing a GUI for that is overkill. When the program is done, the output is way too much to put into a GUI. So most developers think "Why spend time writing a GUI."

Further, a GUI can actually get in the way of running programs efficiently. Most analysis programs have a pretty big footprint and most folks have analysis servers to run on (as opposed to bringing their workstation to it's knees). Getting an xhost environment up and running can be tricky, and a GUI program often doesn't allow you to launch and forget. Personally, my workstation is a laptop and with command-line tools I can launch an analysis on my cluster and then unplug the laptop and head home, something a GUI would prevent.

Programs like CLC-bio provide a nice wrapper around the command-line tools. The downside is they aren't free, and you still need to use the command line unless your analysis is pretty vanilla.

My advice to you is to work hard at getting comfortable with the command line, since you will be using to some extent it for any real bioinformatics research you do.

[simonandrews]

I'd say that more fundamentally there are classes of program which are better suited to a command line environment whilst other functions are better provided through a GUI.

Command line programs tend to be better where you know ahead of time exactly what you want to do so you can set up all of the appropriate options at the start and then leave the program to run. These types of programs lend themselves well to automation and the generation of workflows and pipelines. In terms of sequencing I'd place activities such as base calling, mapping, assembly and SNP detection in this category.

GUI programs tend to be better for exploring data when you want to try out lots of things and get immediate feedback about how well they worked. In sequencing this would be the final visualisation part of the analysis or any other step which involved reviewing and evaluating results. Sometimes these sorts of programs can be useful in helping to decide which parameters to use for subsequent runs of command line programs, but they're really beneficial where you have diverse samples and everything needs to be treated differently.

This isn't to say that you can't write a GUI program to do mapping or review results on a command line, but just that that may not be the optimal way to work.