One of the coolest new uses for mass spectrometry is in the field of protein identification.
One uses a tandem mass spectrometer, in which a mixture containing proteins is first separated (by mass/charge) by the first mass spectrometer. Then, this protein/peptide is fragmented, and individual fragments are detected. This gives you a spectrum that is chracteristic for each unique peptide.
The spectrum alone could be used to arrive at a peptide sequence via brute-force techniques, but the clever bit uses the public DNA/protein database known as Genbank. By using a program to generate theoretical mass/charge spectrums for each peptide in the database, one can match up the measured spectrum with an entry in the database.
I know it works, I got data from it that's going into my thesis.