Soybean meal (SBM) is the preferred protein source in poultry dietary formulations due to its favorable amino acid profile. However, with increasing popularity of all vegetable diets, higher inclusion rates of SBM to satisfy the high protein requirements of turkey poults contributes concerning levels of anti-nutritional factors (ANF) that may have a detrimental effect on health and growth performance of these immature birds. One nutritional strategy to mitigate the negative effects associated with soy-ANFs is to include an alternative vegetable protein source and decrease the amount of SBM included in the diets. An experiment was conducted to evaluate live performance of increasing levels of enzyme-treated soy protein (ESP) in young turkey diets. 480 day-old female Nicholas Select poults were placed in 64 battery pens (7 or 8 birds/pen; 0.5 ft2/bird). A randomized complete block design was utilized, and replicate pens were assigned to one of four dietary treatments: 0%, 5%, 10%, or 20% ESP. Poults were fed a two-phase feeding program consisting of a starter (0-21 days) and grower (21-42 days). Dietary treatments in each phase were formulated to be iso-caloric and contained similar amino acid ratios relative to digestible lysine based off breeder recommendations. All data were analyzed using one-way ANOVA and means deemed significantly different at p≤0.05 were separated using Student’s t-test. Quadratic broken line and quadratic polynomial regressions were used to estimate optimal dose response of ESP. A significant effect of ESP inclusion was observed on FCR at 7 (p=0.011), 28 (p=0.020), 35 (p=0.003), and 42 (p=0.030) days of age. At each of these evaluation periods, a significant reduction in FCR was observed with 10% ESP compared to 0 and 5% with 20% ESP inclusion being intermediate. The quadratic polynomial regression was significant (p=0.04) for BW at day 42, with a breakpoint at 11.52% ESP, whereas the quadratic broken-line regression was not significant (p=0.50). For FCR in the grower period (21-42 days), the quadratic broken-line regression showed a near-significant trend (p=0.09) with a breakpoint of 5.9% ESP, whereas the quadratic polynomial regression did not (p=0.35).