If I want to learn more about how researchers in bioinformatics develop software, a good place for me to start would be to talk to Rachel Brem, whose interaction with software is described as follows (in MCB Fall 2006 MCB Transcript):

To mine data-sets, Brem nearly always writes her own software. "I have a pathology where I don’t like to use other people’s software. There are people who build their own microscopes and other lab hardware. I’m that way with software." It’s a common refrain from bioinformaticists. If you want to understand how the program is treating the data, you had better understand the code. But reading someone else’s code is as challenging as reading someone else’s lab notebook, Brem says. So it’s often faster, not to mention safer, to build your own.