Step 1. READ the manual.
Seriously. You don’t have to read the whole thing in detail but at least understand what the main parameters are, what the program does and what kind of output files to expect.
Most importantly – you need to understand the biological relevance of the output files the tool will generate. Remember that bioinformatics is as much about the biological theory as the (you guessed it) informatics.
Step 2. Explore the test data.
Most tools come with example or test data sets. First look at the test data, it’s size and file formats, look inside the test files and check them out. This way you will see how your input files should be formatted.
Next use this test data try out the most basic example set out in the manual or – if you’re lucky – tutorial.
This will get you familiar with the tool commands and show you what your output should look like. You will need to know what to expect when troubleshooting your analyses if/when they go wrong.
Next, still using the test data, enter in the most extreme values you can think of for different parameters.
Test the limits of what the program can do to get a intuitive understanding of how it works and importantly how changing the parameters might influence your results (again relate this to the biological interpretations).
If you are working on a cluster, using the test data will also give you a useful insight into the expected job time and computational resources needed.
Step 3.Google for help.
Yep, I said it.
If you get stuck, don’t hesitate to google for an answer – this will save you a lot of time. Look for search results that take you to sites like Stack overflow.
It is also handy to look over methods sections of scientific publications for some clues.
Googling error messages is particularly helpful, it usually leads you to a solution that some other poor soul had to work out (one day it may be you, so you can return the favor).
Step 4. Try your own data.
When you’re done experimenting try out the program on your own data – this is often the time you really get a good understanding of a what a program does, no one knows your data like you do 😉 and if in doubt, repeat Step 3.
Good luck!