Smoothed Nested Testing on Directed Acyclic Graphs.

We consider the problem of multiple hypothesis testing when there is a logical nested structure to the hypotheses. When one hypothesis is nested inside another, the outer hypothesis must be false if the inner hypothesis is false. We model the nested structure as a directed acyclic graph, including chain and tree graphs as special cases. Each node in the graph is a hypothesis and rejecting a node requires also rejecting all of its ancestors. We propose a general framework for adjusting node-level test statistics using the known logical constraints. Within this framework, we study a smoothing procedure that combines each node with all of its descendants to form a more powerful statistic. We prove a broad class of smoothing strategies can be used with existing selection procedures to control the familywise error rate, false discovery exceedance rate, or false discovery rate, so long as the original test statistics are independent under the null. When the null statistics are not independent but are derived from positively-correlated normal observations, we prove control for all three error rates when the smoothing method is arithmetic averaging of the observations. Simulations and an application to a real biology dataset demonstrate that smoothing leads to substantial power gains.

Influence of corn hybrid and processing method on nutrient digestibility, finishing performance, and carcass characteristics.

A finishing trial and a digestion trial were conducted to evaluate the effects of corn hybrid and processing method on nutrient digestibility, finishing performance, and carcass characteristics. A 2 x 5 factorial arrangement of treatments was used for both trials. Factors included 2 processing methods, dry-rolled (DRC) or high-moisture (HMC), and 5 commercially available corn hybrids. The finishing trial (Exp. 1) utilized 475 yearling steers (379 +/- 15 kg initial BW), stratified by BW into 2 blocks then assigned randomly to 1 of 60 pens (8 steers/pen). Treatments were assigned randomly to pens with 6 pens/treatment. Diets consisted of 67.5% corn (each hybrid processed as DRC or HMC), 20% wet corn gluten feed, 7.5% alfalfa hay, and 5% supplement. The digestion trial (Exp. 2) used 2 ruminally and duodenally fistulated Holstein steers (560 kg of BW) and the mobile bag technique. Bags were ruminally incubated for 22 h. For total tract digestibility, bags were subjected to a simulated abomasal pepsin digestion, inserted into the duodenum, and collected in the feces. No significant interactions between corn hybrid and processing method were observed in Exp. 1; therefore, only main effects are presented. Feeding HMC decreased (P < 0.01) DMI and increased (P < 0.01) G:F compared with cattle fed DRC. Hybrid tended to affect (P = 0.12) G:F, but did not affect (P > 0.36) any other variable. For Exp. 2, feeding HMC increased (P < 0.01) DM and starch digestibility compared with DRC, but to varying degrees depending on the hybrid. Correlating the results of the 2 trials across processing methods, strong relationships were observed between G:F and postruminal starch digestibility (r = 0.84) and total-tract starch digestibility (r = 0.73). When evaluated within processing method, these relationships weakened and were not significant. The results of these trials indicate that processing method had a larger effect on performance and digestibility than hybrid, and no interaction was observed between processing method and hybrid, suggesting hybrid effects were consistent across processing methods.

The correlation of chemical and physical corn kernel traits with production performance in broiler chickens and laying hens.

A study was conducted to determine the influence on broiler chicken growth and laying hen performance of chemical and physical traits of corn kernels from different hybrids. A total of 720 male 1-d-old Ross-308 broiler chicks were allotted to floor pens in 2 replicated experiments with a randomized complete block design. A total of 240 fifty-two-week-old Hy-Line W-36 laying hens were allotted to cages in a randomized complete block design. Corn-soybean meal diets were formulated for 3 broiler growth phases and one 14-wk-long laying hen phase to be marginally deficient in Lys and TSAA to allow for the detection of differences or correlations attributable to corn kernel chemical or physical traits. The broiler chicken diets were also marginally deficient in Ca and nonphytate P. Within a phase, corn- and soybean-based diets containing equal amounts of 1 of 6 different corn hybrids were formulated. The corn hybrids were selected to vary widely in chemical and physical traits. Feed consumption and BW were recorded for broiler chickens every 2 wk from 0 to 6 wk of age. Egg production was recorded daily, and feed consumption and egg weights were recorded weekly for laying hens between 53 and 67 wk of age. Physical and chemical composition of kernels was correlated with performance measures by multivariate ANOVA. Chemical and physical kernel traits were weakly correlated with performance in broiler chickens from 0 to 2 wk of age (P<0.05, | r |<0.42). However, from 4 to 6 wk of age and 0 to 6 wk of age, only kernel chemical traits were correlated with broiler chicken performance (P<0.05, | r |<0.29). From 53 to 67 wk of age, correlations were observed between both kernel physical and chemical traits and laying hen performance (P<0.05, | r |<0.34). In both experiments, the correlations of performance measures with individual kernel chemical and physical traits for any single kernel trait were not large enough to base corn hybrid selection on for feeding poultry.

The correlation of chemical and physical corn kernel traits with growth performance and carcass characteristics in pigs.

Corn kernel composition may affect its nutritive value and, thus, pig growth performance and carcass characteristics. The objective of this study was to determine the influence of the chemical and physical traits of corn kernels from different hybrids on the growth performance and carcass characteristics of pigs. A total of 288 crossbred pigs were grown in a 3-phase program from 21 kg of BW until slaughter at 113 kg of BW with 12 pens (4 pigs/pen) per dietary treatment. Target BW for each phase were 20 to 40 kg (grower 1), 40 to 80 kg (grower 2), and 80 to 120 kg (finisher). In each phase, diets were formulated to be marginally deficient in Lys, TSAA, Ca, Na, and nonphytate P to improve the likelihood of detecting differences in performance due to corn hybrid. Each of 6 corn hybrids represented a wide range of kernel chemical and physical traits and was substituted for corn in a common diet formulation on an equal weight basis to make the 6 dietary treatments. Physical and chemical composition of the kernels were analyzed and correlated with performance measures by multivariate ANOVA. Kernel density was correlated with i.m. fat (IMF) content in LM (r = -.35, P < 0.05). Stenvert grinding time was correlated (P < 0.05) with ADG during the grower 1 phase (r = 0.26), ADFI during the grower 2 phase (r = 0.27), final BW (r = 0.27), and IMF (r = -0.36). The amylose content of the cornstarch was correlated (P < 0.05) with ADG during the grower 2 phase (r = -0.28) and with BW at the end of the grower 2 phase (r = -0.27). The NDF content of the kernels was correlated (P < 0.05) with ADG during the finisher phase (r = -0.30), final BW (r = -0.33), and number of days to market (r = 0.31). The ADF content of the kernels was correlated (P < 0.05) with ADG during the grower 1 phase (r = -0.26), final BW (r = -0.26), and IMF (r = 0.31). The correlations of performance measure variation with individual kernel hybrid physical and chemical traits were statistically significant yet not large enough to base corn hybrid selection for feeding pigs on any single kernel chemical or physical trait.

Influence of corn hybrid traits on digestibility and the efficiency of gain in feedlot cattle.

Two experiments were conducted to determine the influence of chemical and physical corn kernel traits on digestibility and feedlot cattle performance. Seven commercially available corn hybrids representing a range in kernel traits were evaluated for a variety of chemical and physical traits that included test weight, 1,000-grain weight, kernel size, starch, CP, amylose, Stenvert Hardness tests (kernel hardness traits), tangential abrasive dehulling device loss, 12-h in vitro starch disappearance, and rate and extent of in situ DM disappearance. Differences among hybrids existed for all physical kernel traits measured. In Exp. 1, 224 steers in 28 pens were fed the same hybrids for 167 d in a completely randomized design. All diets were formulated to have 12.5% CP and contained 66% dry-rolled corn (DM basis). There were no differences (P >0.20) among corn hybrids for DMI, ADG, or carcass characteristics. However, efficiency of gain (G:F) was influenced by hybrid (P < 0.01) with a difference of 9.5% from the least to the most efficient. In Exp. 2, 7 ruminally cannulated heifers were used in a 7 x 7 Latin square design to determine the effects of these hybrids on nutrient digestibility, VFA concentrations, and ruminal pH. Total-tract OM and starch digestibilities were not different (P >0.15) among hybrids and averaged 77.9 and 95.1%, respectively. Differences (P < 0.05) among hybrids existed for ruminal propionate concentration and the acetate to propionate ratio. Kernel hardness traits correlated (P < 0.05) with G:F were 1,000-grain weight (r = -0.81), Stenvert time to grind (r = -0.83), and the proportion of Stenvert soft to coarse particle height (r = 0.83). Propionate concentration was not correlated (r = 0.45) with G:F but was correlated (P = 0.02) to the Stenvert time to grind (r = -0.83). Cattle fed dry-rolled corn hybrids with greater proportions of soft endosperm had greater concentrations of propionate and gained more efficiently than cattle fed hybrids with a harder endosperm. Selecting for these softer kernel traits may improve the efficiency of gain in feedlot cattle.