A large number of species cannot be distinguished via standard non genetic analysis in the lab. Here we address the problem of finding minimum sets of restriction enzymes that can be used to unequivocally identify the species of a yeast specimen by analyzing the size of digested DNA fragments in gel electrophoresis experiments. The problem is first mapped into set covering and then solved using Constraint Programming techniques. Although the data sets used are relatively small (23 yeast species and 331 enzymes), a similar approach might be applicable to larger ones and to a number of variants as discussed in the conclusion. One of the set covering algorithms invented gave rise to the need of implementing a variation of the NValues Global Constraint to brake symmetries in the solutions. This thesis will contribute to the expansion of the set of Global Constraints available in the CaSPER system (constraint solving platform for engineering and research).