I think it is true in general that sheafification is left adjoint to inclusion, and commutes with all limits and colimits.
To a sheaf of sets, one can associate a sheaf of abelian groups: Compose with the free abelian group functor, and then sheafify (Voevodsky thesis, p 2). This is the left adjoint to the forgetful functor from abelian sheaves to sheaves of sets. This is an example of a functor from sheaves of sets to sheaves of something else, which is needed if one wants to embed a category into a category behaving like “something else”, in this case an abelian category. For basic properties of these abelian sheaves, see p 2 in the thesis (right exactness, preserves monos, “no torsion”, flatness, respects prod and coprod). More basic properties of these sheaves, in section 2.
http://ncatlab.org/nlab/show/sheafification
http://mathoverflow.net/questions/51467/naturality-of-the-associated-sheaf
http://www.ncatlab.org/nlab/show/local+isomorphism
http://ncatlab.org/nlab/show/abelian+sheaf
http://ncatlab.org/nlab/show/quasicoherent+sheaf
http://ncatlab.org/nlab/show/IPC-property
http://www.ncatlab.org/nlab/show/restriction+and+extension+of+sheaves
http://ncatlab.org/nlab/show/category+of+sheaves
http://www.ncatlab.org/nlab/show/local+epimorphism
http://ncatlab.org/nlab/show/fine+sheaf
http://www.ncatlab.org/nlab/show/Q-category
http://golem.ph.utexas.edu/category/2010/02/sheaves_do_not_belong_to_algeb.html
http://mathoverflow.net/questions/2314/several-topos-theory-questions related to the functor of points viewpoint and more
Any presheaf is a colimit of representables. Can you say filtered here??
nLab page on Sheaf theory