Two laboratories independently examined how the amount of starting material influences DNA extraction efficiency and, ultimately, the detection of low-level presence of genetically engineered (GE) traits in commercialized grains. GE traits from one maize, two canola, and two soybean samples were used as prototypical models in the study design as well as two commonly used DNA extraction methods, a small scale (0.1 and 0.2 g samples) and a large scale (1.0 and 2.0 g samples). The DNA samples were fortified (spiked) at 0.1 and 0.01% (w/w) levels. The amount of DNA recovery varied between the two laboratories, although a sufficient amount of DNA was obtained to perform replicate PCR analysis by both laboratories. Reliable detection of all five events was achieved by both laboratories at 0.1% level using either small-scale or large-scale DNA extractions. Reliable detection of the GE events was achieved at 0.01% level for soybean and canola but not for maize. Variability was observed among the two laboratories in terms of the Ct values generated. There was no difference between small-scale and large-scale DNA extraction methods for qualitative PCR detections of all five GE events.