The Plackett-Burman design: An effective vehicle to evaluate multiple variables in poultry nutrition

Abstract

Plackett-Burman (PB) designs are effective as they permit the assessment of 11 factors in a single experiment. The PB design ranks factors by their impact on response parameters, which is valuable for identifying factors to be optimized in subsequent studies. This design is novel to animal studies; therefore, the objective of this study was to evaluate the PB design in the context of phytase supplemented broiler diets. Eleven dietary factors were screened over 12 treatments at 2 levels (canola meal [0 or 125g/kg], wheat or maize [600 g/kg], whole barley [0 or 200 g/kg], digestible lysine [9.6 or 11.4 g/kg], phytate-P [2.6 or 3.3 g/ kg], Ca [6 or 10 g/kg], available P [3.0 or 4.5 g/kg], Na [1.6 or 2.0 g/ kg], xylanase level 1 [0 or 1200 U/kg], xylanase level 2 [0 or 2000 U/ kg], xylanase level 3 [0 or 2000 U/kg]; 3 xylanase factors were used to create a titration of xylanase activities derived from Trichoderma reesei of 0 to 5200 U/kg). Steam-pelleted diets based on wheat/maize and soybean meal with 1000 FTU/kg Buttiauxella phytase were offered to 468 male Ross 308 chicks (6 birds/cage, 6 replicates/treatment) from 7 to 28 d post-hatch. The effects of dietary treatments on weight gain, feed intake, gain:feed, relative gizzard weights, contents and pH were determined. Data were analyzed in JMP Pro 13 (SAS Institute Inc. JMP Software, Cary, NC) via PB screening methodology, with significant differences when P ≤ 0.05. Digestible lysine level was the most important positive factor influencing weight gain, feed intake and gain:feed (P < 0.001). The high digestible lysine level increased gain by 10.2% (1228 versus 1419), intake by 4.92% (1889 versus 1982) and gain:feed by 10.3% (0.649 versus 0.716) compared with the low level. Ca was the most important negative factor influencing growth performance (P < 0.001). The high Ca level depressed gain by 6.36% (1367 versus 1280), intake by 3.85% (1974 versus 1898) and gain:feed by 2.60% (0.692 versus 0.0.674) compared with the low level. Whole barley was the most important factor positively influencing gizzard weights (P < 0.001) and gizzard contents (P < 0.006); the 200 g/kg inclusion increased gizzard weights by 24.8% (19.43 versus 24.25) and contents by 16.0% (7.68 versus 8.91) compared with ground grain diets. Ca was the most important factor influencing gizzard pH (P < 0.002) where the high Ca level increased pH (2.78 versus 3.01) compared with the low level. In conclusion, the PB design is a novel, effective vehicle to assess multiple factors in feeding studies for poultry.