Myelodysplastic syndrome (MDS) is a clonal disorder of hematopoietic stem cells characterized by ineffective and inadequate hematopoiesis. MDS is also a susceptibility to acute myeloid leukemia (AML) and shown to be extremely resistant to current therapeutic strategies. MDS in a subset of 10-20% of patients arise after previous chemotherapy or radiation exposure for other malignancies. Because MDS is a heterogeneous disorder, specific gene abnormalities playing a role in the myelodysplastic process have been difficult to identify. Cytogenetic abnormalities are seen in half of MDS patients, and generally consist of partial or complete chromosome deletion or addition, whereas balanced translocations are rare. Genes more frequently implicated in the pathogenesis of MDS remain unknown. Although point mutations of critical genes have been demonstrated to contribute to the development MDS, there was no strong correlation between these mutations and clinical features. Recently, we reported the high incidence of somatic mutations in the AML1/RUNX1 gene, which is a critical regulator of definitive hematopoiesis and the most frequent target for translocation of AML, in MDS, especially refractory anemia with excess blasts (RAEB), RAEB in transformation (RAEBt) and AML following MDS (defined here as MDS/AML). The MDS/AML patients with AML1 mutations had a significantly worse prognosis than those without AML1 mutations. Most of AML1/RUNX1 mutants lose trans-activation potential, which leads to a loss of AML1 function indicating that AML1/RUNX1 dysfunction is one of the major pathogenesis of MDS/AML. Normalizing AML1 function or regulating cooperative gene mutations would provide an important clue for molecular target therapies.