The fragments depend on the ionization technique you use. Usually the recent literature/instrument manuals or software provide a table for m/z and corresponding possibilities of chemical identities (including fragments). But again it is specific to the mass spectrometeric technique you use.
There is no single rule that governs the fragmentation of molecules in Mass Spectrometry. The same molecule can give different fragments in the same Mass Spectrometer but under different conditions. It is not possible to predict what fragmentation pattern can happen to a molecule in MS, especially if the molecule is too big (with molecular weight higher than 400 for instance) and has many branches.
Usually the software that runs the MS has the ability to detect the different fragments that result from the fragmentation of the parent molecule after being ionized (parent ion). That software should be able to program the MS in a specific way to put the molecule under certain ionization conditions and collision energy and then to detect the fragments.
To develop a MRM or a SRM method, you need to consult your software and see how can you do so. Our instrument is Waters, USA. The software is Masslynx, and the part of the software that develops the methods is called IntelliStart.
Just want to add to the answers above - if you are considering predicting peptide fragmentation, you can use the following software
http://www.gpmaw.com/
If you type the sequence of the petide in the utility called "graphical fragment mapper", you will be able to predict whatever N- or C-terminal directed fragments you need to see.
If you are considering prediction of fragemnts of a compound that is not a peptide, that is more difficult to do.
mannich base produce from two compounds of secondary amine with aldehyde .... but in most of papers we found SH , OH react with aldehyde and ketone , which one is the right ?? or ( R)2N with ketone and aldehyde is mannich reaction ??which one from these (( 3 types ) act mannich reaction?
ChemSketch and Chemtool (http://ruby.chemie.uni-freiburg.de/~martin/chemtool/chemtool.html) for structures of small molecules and Avogadro ( http://avogadro.cc/wiki/Main_Page ). You can download proteins directly from the RCSB PDB ( http://www.rcsb.org/pdb/home/home.do ) and online java based drawing module at http://eawag-bbd.ethz.ch/predict/
For general fragmentation in a mass spectrometer there are some general rules for molecules. The fragments formed will depend on the energy of your ionization technique and the amount of energy it takes to break the various bonds in your molecule. So you can predict them to some extent, although as molecules get larger, prediction gets more complex, and you sometimes get rearrangement of the molecule or group before fragmentation which makes it more complicated still. You will also get 'secondary fragmentation' where a large fragment of a molecule fragments into smaller groups and so on. Try a book on mass spectrometry - there are even books on specific areas of mass spectrometry (proteins, polymers, lipids, different ionization techniques etc.). "Interpretation of mass spectra" by McClafferty is the classic text but there are others which will give you a good introduction to the basics.