Functional dependence graphs are directed acyclic graphs where the
  nodes represent variables, and the directed edges represent
  functional dependence.  The concept of fd-separation introduced by
  Kramer extends d-separation to functional dependence graphs. It is
  of interest to determine all sets of variables which fd-separate two
  other sets of variables in an FDG. This corersponds to deterining
  all conditional independence relationships compatible with the
  graph. Acid and de Campos found an algorithm for finding
  d-separating sets. In this note, we extend theor work to provide a
  method for determining minimal fd-separators.