The best tool to learn depends on the actual job. For very basic data analysis, Excel has the lowest entry barrier, but the learning curve gets steep soon. Learning a full language will take some months, but will greatly expand the things she is capable of doing. Also, knowing programming in a field where everybody uses Excel can be a great advantage and unique feature for employment.
One important factor deciding which language is the environment she is in. Having people used to the tools and perks of a language can be very useful when learning from scratch. On the other hand, I'd only count "modern" options (some old professors are very fond of FORTRAN77 and IDL, but learning these is like learning to train dinosaurs).
She should consider not only her lab, but also ask at the Bioinformatics department, and take a look at other labs in the branch she is in. In my experience, for the most informatics side of the field, people use mostly Python, C++, and Java for some machine learning. I have encountered some MATLAB, but definitely not much.
One last note: I think in learning your first language you want the cleanest and less quirky possible. MATLAB in this respect is a mess of a language, with a crappy syntax and bad scalability. My personal choice would be Python, so here are some introductory materials:
- Summer school on scientific programming and introduction to Python. All materials are posted.
- The official tutorial. There is a lot of things that are not for useful for scientists, but it is very good introduction to the language.
- Khan academy. They have everything. Very nice.
- Coursera has a full assortment of courses. She will surely find something of interest there.