The Food and Drug Administration's (FDA's) 1992 policy statement granted genetically engineered foods presumptive GRAS (generally recognized as safe) status. Since then, divergent views have been expressed concerning the scientific support for this policy. This paper examines four sources to better understand the basis for these claims: 1) internal FDA correspondence; 2) reports from the National Academy of Sciences; 3) research funded by US Department of Agriculture from 1981 to 2002; and 4) FDA's proposed rules issued in 2001. These sources reveal that little research has been conducted on unintended compositional changes from genetic engineering. Profiling techniques now make this feasible, but the new debate centers on the functional meaning of compositional changes.