Modeling improvements in digestible amino acids by a consensus bacterial 6-phytase variant in grower pigs

Abstract

Data from 8 datasets generated from 5 independent experiments that determined the effects of a consensus bacterial 6-phytase variant (PhyG) on apparent ileal digestibility (AID) of amino acids (AA) in growing pigs (~21 to 45 kg body weight) were combined and modeled to test the hypothesis that the phytase results in significant improvements in ileal AA digestibility. The aim was to generate accurate and robust dose-related predictions of the digestible AA contributions of the phytase. The 5 experiments were conducted in Spain, Australia, USA, and Brazil and incorporated variation in diet composition (ingredient composition, phytate-phosphorus (PP) content, limestone solubility), diet form, animal breed and sex. A total of 325 datapoints (observations) were analyzed. First, the relationship between the percentage AID of AA and log-transformed phytase dose (range 0 to 4,000 phytase units (FTU)/kg, analyzed values) was modeled by log-linear regression across all datasets without adjustment for variation in the response to the negative control (NC) diet. The model predicted that the mean AID of total AA in the NC diets was 68.6% (range 43.3% [Cys] to 81.7% [Arg]). This was increased linearly by PhyG supplementation (P < 0.05), by 3.7 percentage units when dosed at 1,000 FTU/kg and by 4.5 percentage units at 4,000 FTU/kg. Second, the percentage unit change in the AID of AA at each phytase dose from baseline (NC without added phytase) was calculated separately for each dataset and the data then combined and modeled together by log-linear regression, against analyzed and log-transformed phytase dose. By this analysis, increases were evident for the AID of all individual (and total) AA. Increases (vs. baseline) at 1,000 and 2,000 FTU/kg were greatest for Trp (+ 6.1 and + 6.8 percentage units), Thr (+ 3.4 and + 3.8 percentage units after correction for synthetic Thr), Gly (+ 7.5 and + 8.3 percentage units) and Cys (+ 5.6 and + 6.2 percentage units). In conclusion, combined data from 5 separate experiments indicate that the bacterial phytase-6 variant will improve ileal digestibility of AA if included in diets fed to growing pigs. The data will allow diet-specific AA matrix recommendations to be made in commercial feed formulations containing PhyG.

Modeling data from 8 datasets generated from 5 independent experiments testing the effect of the same phytase (PhyG) demonstrated that the phytase increases amino acid digestibility in growing pigs under a range of production settings. The results support the use of a digestible AA matrix for PhyG in pig diets.

Introduction

Pig feed ingredient prices remain high, particularly for high-quality protein sources such as soybean meal. Producers formulate diets on a least-cost basis but are also increasingly concerned with reducing the environmental impact of production as part of efforts to improve sustainability. Maximizing feed efficiency to reduce P and nitrogen excretion is an important part of achieving sustainability goals. Hence, strategies that increase the ileal digestibility and retention of dietary proteins and amino acids (AA) are of major interest to producers.

Exogenous microbial phytases are routinely added to pig diets to improve P digestibility and retention (Selle and Ravindran, 2007; Humer et al. 2015). Phytase effects the stepwise dephosphorylation of plant-derived phytate to release pyrophosphate which can be absorbed and utilized by the animal (Greiner and Konietzny, 2011). In degrading phytate, phytase also improves the digestibility of protein and AA (Adedokun et al., 2015; Espinosa et al., 2022; Lagos et al., 2022). This is thought to result primarily from a reduction in the formation of binary- (protein-phytate) and ternary- (protein-phytate-Ca²+) complexes in specific regions of the gastrointestinal tract due to the reduced abundance of phytate (Selle et al., 2012). Phytase also reduces mucin secretion (Selle et al., 2012) which results in reduced endogenous AA flows (Cowieson et al., 2004), increased intestinal uptake of AA via Na+, K+-ATPase pump activity (Liu et al., 2008) and increased sodium concentrations within enterocytes, all of which may also contribute to improved AA retention.

The greater inherent capacity for nutrient digestion of pigs compared with poultry may lead to the expectation that phytase would have a lower effect on protein and AA digestibility in pigs than in poultry. Nevertheless, in a meta-analysis of 28 experiments, Cowieson et al. (2017) reported that phytases dosed at 250 to 20,000 FTU/kg improved the apparent ileal digestibility (AID) of AA by, on average, 2.8 percentage units compared with an unsupplemented diet. In practice, the actual degree of improvement achieved varies between individual phytases and in different settings. For example, one study reported that a Buttiauxella sp. phytase included at 1,046 phytase units (FTU) per kg of a corn-soybean meal-based diet increased the AID of total AA by 3.7 percentage units compared with an unsupplemented diet, whereas a comparator Escherichia coli phytase at equivalent dose had no effect on the AID of AA when tested under the same conditions (Dersjant-Li and Kwakernaak, 2019). Variability in protein and AA digestibility responses to phytase is dependent not only on phytase dose (Dersjant-Li et al., 2022a; Espinosa et al., 2022) but also on the individual enzymatic and biochemical properties of the phytase (Menezes-Blackburn et al., 2015) and on several dietary factors including the solubility of the limestone used in the diet (Bello et al., 2022a; Velayudhan et al., 2022), the feedstuff composition, the AA and protein sources used, their concentrations and inherent digestibility (Ravindran et al., 1999; Cowieson et al., 2017), the concentration of phytate present (Ravindran et al., 1999; Babatunde et al., 2021), and animal-related factors such as genetics (Li et al., 2015) and age (Lagos et al., 2022). This variability means that digestible AA “matrix values” (AA reductions that are applied to the diet to account for the expected contribution of the phytase) need to be derived on an individual phytase basis and generated from multiple studies that encompass as many of the factors that influence the response as possible. The supporting data from such studies and how they can be combined to generate a robust matrix are rarely published. The present analysis sought to address this for one particular phytase.

Ileal AA digestibility responses of growing pigs to phytase inclusion at different levels were modeled against phytase dose using data from 8 datasets generated from 5 digestibility experiments to generate digestible AA matrix values. The primary hypothesis was that a dose-dependent increase in AID of AA is observed if the phytase is included in diets fed to growing pigs. A secondary hypothesis was that the generated model could be used to illustrate how a digestible AA matrix for a phytase can be derived for a diet of known AA concentration.

Materials and Methods

Ethical considerations

Experiment 1 was conducted in accordance with European Directive 2010/73/EU and the Spanish guidelines for the care and use of animals in research (B.O.E. number 252, Real Decreto 2010/2005). Experiment 2 was conducted in accordance with the Queensland Animal Care and Protection Act (2001) and the Australian Code for the Care and Use of Animals for Scientific Purposes (2013). Protocols used in experiments 3 and 4 were approved by the Institutional Animal Care and Use Committee at the University of Illinois, Urbana, USA. Experiment 5 was approved by the Ethics Committee on Animal Use of the Federal University of Lavras, Brazil.

Overview of experiments

Data from 5 experiments conducted in 4 geographical locations were included in the analysis. The main features of the included experiments are detailed in Table 1. The 5 experiments generated 8 datasets, with dataset being defined as the subset of datapoints obtained from an experiment that related to basal diets of the same ingredient composition. In total, 325 data points (experimental units) were included in the analysis.

Table 1.

Overview of experiments¹

Experiment No..2	Dataset No.	Breed	Sex	BW at start of trial (at ileal digesta sampling), kg	No. treatments3	No. replicates per treatment	Phytase (PhyG) dose levels, FTU/kg	Dietary phytate-P (PP) level, %4	Total Ca level, %4	Solubility of Ca source (limestone), %5	Total P level, %4	Diet form	Main ingredients6
Experiment 1	1	Pietrain x (Large White × Landrace)	mixed sex	6.0 (~28)	4	9	0, 250, 500, 1,000	0.27	0.50	53	0.40	mash	Corn, SBM, RB
Experiment 2	2	Large White	males	12.8 (~28)	5	8	0, 250, 500, 1,000, 2,000	0.30	0.59	70	0.45	pellet	Wheat, corn, SBM, RM, RB
	3								0.59	92	0.45
Experiment 3	4	PIC	mixed sex	12.7 (~24)	5	8	0, 500, 1,000, 2,000, 4,000	0.23	0.55	62	0.37	mash	Corn, SBM, RSM
	5							0.29	0.55		0.43
	6							0.35	0.55		0.49
Experiment 4	7	PIC	males	17.8 (~22–45)7	6	9	0, 250, 500, 1,000, 2,000, 4,000	0.23	0.53	62	0.35	mash	Corn, SBM, RSM
Experiment 5	8	DB GENETIC	mixed sex	6.5 (~21)	5	7	0, 250, 500, 1,000, 2,000	0.24	0.58	87	0.36	mash	Corn, SBM, WB

¹In total, 8 datasets and 325 observations.

²Experiment 1: Institut de Recerca I Tecnologia Agroalimentàries, Constantí, Spain; Experiment 2: The University of Queensland, Gatton, Australia; Experiment 3 & 4: University of Illinois, Urbana, IL, USA. Experiment 5: Animalnutri Ciência e Tecnologia, Lavras-MG, Brazil.

³Additional treatments concerning other phytases or positive control diets were included in trials 1, 3 and 5. These were excluded from the present analysis.

⁴Calculated composition, averaged across NC and NC + PhyG treatments.

⁵Determined after 5 min incubation at pH 3.0 and 42 °C, according to the method of Kim et al. (2019).

⁶SBM, soybean meal; RSM, rapeseed meal; PR, polished rice; RB, rice bran; WB, wheat bran.

⁷Digesta collected via cannula over a 2-d period.

BW, body weight; PIC, Pig Improvement Company (Line 359 boars and Camborough females).

All experiments used a randomized complete block design except Experiment 4, which used an incomplete Latin square design with 6 diets and three 11-d periods. All experiments were designed to test the effect of a consensus bacterial 6-phytase variant (herein termed PhyG; Danisco Animal Nutrition & Health (IFF), Oegstgeest, the Netherlands) on the AID of AA. Effects on nutrients other than AA and other response measures are reported elsewhere. The composition of the experimental diets is given in Supplementary Tables S1 to S5.

Phytase enzyme, diet preparation, and provision

The test phytase was a consensus bacterial 6-phytase variant produced in Trichoderma reesei. The safety, biochemical and enzymatic characteristics of this phytase have been described (Christensen et al., 2020; Ladics et al., 2020). The phytase was premixed with wheat (experiments 1 and 2), or ground corn (experiments 3, 4, and 5) prior to its addition to the experimental diets. All final diets were thoroughly mixed to ensure a homogeneous distribution of phytase and other ingredients. Diets were provided ad libitum in experiments 1, 2, 3 and 5. In experiment 4, experimental diets were provided at a daily rate of 4.5% of body weight (BW) recorded at the beginning of each period. Water was freely available in all experiments.

Common chemical analyses

In all experiments, phytate-P (PP) concentrations in the final diets were determined at Danisco Animal Nutrition Research Centre (Brabrand, Denmark) using the high-performance liquid chromatography (HPLC) method described by Christensen et al. (2020), modified from that of Skoglund et al. (1998). Phytase activities in experimental diets were analyzed by Danisco Animal Nutrition Research Centre according to a modified version of AOAC method 2000.12 (Engelen et al., 2001), where one FTU was defined as the quantity of enzyme that released 1 µmol of inorganic phosphate from a 0.0051 mol/L sodium phytate substrate per minute at pH 5.5 at 37 °C.

Experiment 1

This experiment was conducted at the Institut de Recerca I Tecnologia Agroalimentàries (IRTA), Constantí, Spain. Seventy-two crossed Pietrain × (Large White × Landrace) weaned pigs (50% males, 50% females; initial BW 6.0 ± 1.0 kg) were fed a common pre-starter diet until 42 d of age (BW 10 to 11 kg). Pigs were then blocked based on BW and sex and allotted to floor pens (2 pigs per pen, 9 pens per treatment). From 42 to 70 d of age, pigs were fed experimental diets based on corn, soybean, meal and rice bran, supplemented with phytase according to treatment (negative control [NC], NC + PhyG at 250 FTU/kg, NC + PhyG at 500 FTU/kg, or NC + PhyG at 1,000 FTU/kg). The unsupplemented NC diet was formulated to contain 0.50% Ca, 0.11% apparent total tract digestible (ATTD) P (no feed phosphate was added to the diet), 0.27% PP, 1.23% standardized ileal digestible (SID) Lys, and 2,520 kcal/kg net energy (NE). Titanium dioxide was added to all diets as a digestibility marker, at a level of 0.50%. Diets were provided in mash form. Pigs were housed in an environmentally controlled barn where temperature was controlled initially at 32 °C and then reduced by 1 °C per week until 24 °C was reached.

At 70 d of age (after 28 d on trial), one piglet per pen was euthanized by intravenous injection of sodium pentobarbital and the digesta was collected from the distal half of the ileum, with ileum being defined as the length of the small intestine from Meckel’s diverticulum to 4 cm anterior of the ileocecal junction. Samples were homogenized, freeze-dried and stored at − 20 °C until analysis. For digestibility analyses, all digesta samples were freeze-dried and finely ground, prior to analysis. Calcium and P in diets were determined on ashed samples by inductively coupled plasma mass spectrometry (ICP-MS; Agilent Technologies model 7700X) at SCT laboratories (University of Lleida, Spain) according to the method of Pacquette et al. (2018). Amino acids in diets and ileal digesta were determined by acid hydrolysis followed by HPLC determination with pre-column derivatization, according to the procedures of Llames and Fontaine (1994). A prior oxidation in performic acid was required for the determination of sulfur AA. Tryptophan was determined under alkaline conditions (because it is destroyed during acid hydrolysis) according to the method of Delahaye and Landry (1992). Titanium dioxide in feed and digesta were determined according to the method of Short et al. (1996).

Experiment 2

Experiment 2 was conducted at The University of Queensland, Gatton, Australia. A total of 80 Large White pigs (intact males; initial BW 12.8 ± 1.33 kg) were blocked based on BW into individual floor pens (8 pens per treatment) and fed experimental diets for 20 d. Two NC diets were formulated based on corn, wheat, soybean meal, and rapeseed meal. One of the NC diets contained a limestone of medium solubility (70% soluble after 5 min incubation at pH 3.0 and 42 °C according to the method of Kim et al. (2019) and the other NC diet contained limestone of high solubility (92% soluble after 5 min incubation under the same conditions using the same method). Both diets were formulated to contain 0.59% Ca, 0.18% standardized total tract digestible (STTD) P (with no added feed phosphate), 0.30% PP, 1.14% SID Lys, and 2,400 kcal/kg NE. The basal diets were supplemented with phytase according to treatment (NC, NC + PhyG at 250 FTU/kg, NC + PhyG at 500 FTU/kg, NC + PhyG at 1,000 FTU/kg, or NC + PhyG at 2,000 FTU/kg). Titanium dioxide was added to all diets at a level of 0.50%. Diets were provided in pelleted form. Pigs were housed in an environmentally controlled animal facility in which temperature was maintained at 24 to 27 °C. The lighting regime was based on a 16:8 h light:dark cycle.

On d 21 of the experiment, all pigs were euthanized by pentobarbitone sodium injection and ileal digesta was collected. Digesta samples were homogenized, freeze-dried, and stored at − 20 °C for later analysis. Calcium and P in diets were determined by inductively coupled plasma-optical emission spectroscopy (ICP-OES) following microwave-assisted acid digestion, according to the AOAC method 985.01 (AOAC, 1996). Titanium in diet and ileal digesta samples was determined according to the method of Lomer et al. (2000) and measured on ash samples by ICP-MS at the University of Queensland, Australia. Samples for AA analysis were prepared using 24-h hydrolysis in 6 N hydrochloric acid at 110 °C under an atmosphere of nitrogen. Samples were oxidized in performic acid before acid hydrolysis for Met and Cys analyses. Samples for Trp analysis were hydrolyzed using barium hydroxide. Amino acids in hydrolysates were determined by cation-exchange chromatography coupled with post-column derivatization (AOAC, 2000; method 982.30 E [a, b, c]).

Experiment 3

Experiment 3 was conducted at the University of Illinois, Urbana IL, USA. A total of 120 weaned pigs (offspring of Line 359 boars and Camborough females, Pig Improvement Company, Hendersonville, TN, USA; initial BW 12.70 ± 4.01 kg) were blocked based on weaning group into individual floor pens and fed experimental diets for 20 d. Three NC diets were formulated based on ground corn, soybean meal, rapeseed meal and cornstarch, to contain three different PP levels (0.23%, 0.29%, and 0.35%) that spanned the range of levels used in typical commercial diets. These diets were all formulated to contain 0.17 % STTD P (this resulted in no feed phosphates being added to the diet containing 0.35% PP), 0.55% Ca, 1.10% SID Lys and, 2,289 to 2,381 kcal/kg NE. The NC diets were supplemented with phytase according to treatment (NC, NC + PhyG at 500 FTU/kg, NC + PhyG at 1,000 FTU/kg, NC + PhyG at 2,000 FTU/kg, and NC + PhyG at 4,000 FTU/kg). Titanium dioxide was added to all diets at a level of 0.40%. Diets were provided in mash form. Pigs were housed in an environmentally controlled animal facility in which temperature was maintained at 24 to 27 °C with a light:dark cycle of 16:8 h.

On day 20, pigs were euthanized by injection with pentobarbitone sodium and ileal digesta were collected. Digesta samples were homogenized, freeze-dried, and stored at − 20 °C for later analysis. Calcium and P in diets were determined by ICP-OES using an internally validated method (method 985.01 A, B, and C; AOAC, 2007) after wet ash sample preparation (method 975.03 B[b]; AOAC, 2007). Amino acids in diets and digesta were analyzed on a Hitachi amino acid analyzer (Model No. L8800; Hitachi High Technologies America, Inc., Pleasanton, CA, USA) using ninhydrin for post-column derivatization and norleucine as the internal standard (Method 982.30 [a, b, c]; AOAC, 2019). Titanium in diet and digesta samples was determined according to Myers et al. (2004).

Experiment 4

Experiment 4 was also conducted at the University of Illinois, Urbana, IL, USA. A total of 18 castrated male pigs (offspring of Line 359 boars and Camborough females, Pig Improvement Company, Hendersonville, TN, USA; initial BW 17.81 ± 1.71 kg) were fitted with a T-cannula in the distal ileum and allotted to floor pens in a 6 × 3 incomplete Latin square design (6 experimental diets, three 11-d experimental periods, 9 replicate pigs per treatment). The NC diet was based on ground corn, soybean meal and rapeseed meal, and contained no added feed phosphate. This diet was formulated to contain ~0.23% PP, 0.53% Ca, 0.15% STTD P, 1.10% SID Lys, and 2,472 kcal/kg NE. The NC was supplemented with phytase according to treatment (NC, NC + PhyG at 250 FTU/kg, NC + PhyG at 500 FTU/kg, NC + PhyG at 1,000 FTU/kg, NC + PhyG at 2,000 FTU/kg and NC + PhyG at 4,000 FTU/kg). Titanium dioxide was added to all diets at a level of 0.40%. Diets were provided in mash form.

The initial 5 d of each 11-d experimental period was considered the adaptation period and ileal digesta was collected on d 10 and 11 using standard procedures. Digesta samples were homogenized, freeze-dried, and stored at − 20 °C for later analysis. Calcium, P, AA, and titanium in diets and ileal digesta were determined as described for Experiment 3.

Experiment 5

Experiment 5 was conducted at Animalnutri Ciência e Tecnologia, Lavras-MG, Brazil. A total of 500 mixed-sex weaned pigs (DB GENETIC, Patos de Minas, Minas Gerais, Brazil) of approximately 23 d of age (initial BW 6.53 ± 0.89 kg) were blocked based on initial BW and allotted to 5 dietary treatments (10 pens per treatment, 5 males and 5 females per pen) and fed experimental diets for 42 d. The NC diet was based on corn, soybean meal, and wheat bran. This diet was formulated to contain 0.58% Ca, 0.15% STTD P, 0.24% PP, 1.21% SID Lys, and 2,393 kcal/kg NE. The NC was supplemented with phytase according to treatment (NC, NC + PhyG at 250 FTU/kg, NC + PhyG at 500 FTU/kg, NC + PhyG at 1,000 FTU/kg, and NC + PhyG at 2,000 FTU/kg). Titanium dioxide was added to all diets at a level of 0.50%. Diets were provided in mash form. Pigs were housed in an environmentally controlled facility in which temperature was maintained at 24 to 27 °C and lighting was provided for 16 h per d.

On day 42, 2 pigs per pen (1:1 male to female ratio, of average pen BW) from 7 replicate pens per treatment (selected at random) were euthanized and ileal digesta was collected. Digesta samples (pooled per pen) were homogenized, freeze-dried, and stored at − 20 °C for later analysis. Diets were analyzed for Ca (method 968.08; AOAC, 2006), and P (method 946.06; AOAC, 2006). Amino acids were determined using an AA-analyzer (LKB 4151 Alpha plus AA analyzer, LKB Biochrom, Cambridge, UK) according to method 982.30 [a, b, c]; AOAC, 1990. Titanium in diets and ileal digesta was analyzed according to Myers et al. (2004).

Calculations

Values for AID (expressed as percentages) were calculated according to the following equation (Stein et al., 2007):

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{A}\mathrm{I}\mathrm{D}\left(\mathrm{\%}\right)=(1-\left[\mathrm{A}{\mathrm{A}}_{{}_{\mathrm{d}\mathrm{i}\mathrm{g}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{a}}}/\mathrm{A}{\mathrm{A}}_{{}_{\mathrm{d}\mathrm{i}\mathrm{e}\mathrm{t}}}\right]\times \left[\mathrm{T}{\mathrm{i}}_{\mathrm{d}\mathrm{i}\mathrm{e}\mathrm{t}}/\mathrm{T}{\mathrm{i}}_{\mathrm{d}\mathrm{i}\mathrm{g}\mathrm{e}\mathrm{s}\mathrm{t}\mathrm{a}}\right])\times 100}\end{array}}$$

where AA_digesta is the concentration of the AA in the digesta (mg/kg DM), AA_diet is the concentration of the AA in the diet (mg/kg DM), Ti_diet is the titanium concentration in the diet (mg/kg DM), and Ti_digesta is the titanium concentration in the digesta (mg/kg DM).

For determination of the AA digestibility improvement from the supplemental phytase that was achieved above the unsupplemented NC diet, a correction was made to remove the dietary content of synthetic AA (feed-grade AA) from the calculation. This was on the basis that the digestibility of synthetic AA (Lys, Met and Thr) is already close to 100 % (Izquierdo et al., 1998; Oliveira et al., 2020; Selle et al., 2020) and therefore considered unlikely to be improved by supplemental phytase. Its removal from the calculation was expected to yield a more accurate assessment of the degree of improvement in the AA digestibility of raw materials achieved by the phytase.

Statistical modeling

Pig was the experimental unit for all analyses. Relationships between the AID of AA (individual and total AA, as percentages) and increasing phytase dose (based on analyzed values, log-transformed) were modeled in two ways based on the combined dataset from all 5 experiments. First, the AID of AA (percentages) among all datasets was modeled against increasing (analyzed) phytase dose using log-linear regression. Other model types including exponential were tested but the log linear regression was determined to be the model of best fit, as determined by the Akaike Information Criterion. This approach included all the variation among experiments as potential influencing factors. Second, the digestibility response to the NC diet was accounted for by calculating the percentage unit increase in AID of AA above the response to the NC diet, at each analyzed phytase dose, for each dataset separately. The relationship between the percentage unit increase in the apparent ileal digestibility (AID) of amino acids (both individual and total, expressed in percentage points) and the analyzed phytate-phosphorus content of the diet was evaluated; however, no significant correlation was observed (P > 0.05) The data from all datasets were then modeled together against increasing log-transformed analyzed phytase dose using log-linear regression. The aim of the second modeling approach was to gain a truer indication of the degree of improvement achieved by the phytase unimpeded by inherent differences in the AA digestibility of the NC diets. Prior to performing the linear regressions, the analyzed phytase dose data were log-transformed. Data were then back-transformed to the corresponding normal values afterwards. This approach was used because the log-linear model was the model of best fit for the AID of AA response.

All statistical analyses were conducted in JMP16.0 (JMP, 2022). Differences and effects were considered significant at P < 0.05, whereas 0.05 ≤ P < 0.1 was considered to be a tendency.

Results

Analyzed nutrients and phytase in the diets

The analyzed concentrations of nutrients in the diets (averaged across NC and phytase-supplemented diets) are presented for each dataset in Table 2. Within all datasets, analyzed PP concentrations were within 10% of formulated values. Levels of native phytase in the unsupplemented NC diets were generally below 250 FTU/kg except in experiment 2 where analyzed phytase activities were 347 and 326 FTU/kg in the two NC diets containing medium and high soluble limestone, respectively. After deduction of native phytase activities in the respective NC, analyzed phytase activities in the PhyG-supplemented diets confirmed the supplementation of the phytase. Despite some variation between analyzed and formulated phytase activity, good separation between adjacent phytase levels was achieved within all datasets. Analyzed phytase activities were used in the modeling of AA digestibility responses to improve the accuracy of the predictions.

Table 2.

Analyzed nutrient composition (average of negative control and phytase-supplemented diets, % as is, unless otherwise indicated) and phytase activities of the experimental diets in experiments 1 to 5 (8 datasets)

Dataset No.	Approx. BW at sampling, kg	Limestone solubility at 5 min, %	Ca (total)	P (total)	Ca:P (total)	Phytate-P	Arg (total)	His (total)	Ile (total)	Leu (total)	Lys (total)	Met (total)	Phe (total)	Thr (total)	Trp (total)	Val (total)
1	28	53	0.64	0.46	1.35	0.26	1.21	0.36	0.68	1.48	1.30	1.19	0.87	0.84	0.28	0.74
2	28	70	0.74	0.49	1.51	0.31	1.21	0.55	0.89	1.57	1.25	0.31	1.01	0.88	n.a.	1.02
3	28	92	0.75	0.49	1.53	0.29	1.23	0.55	0.90	1.58	1.24	0.35	1.02	0.90	n.a.	1.02
4	24	62	0.61	0.33	1.97	0.22	1.01	0.41	0.71	1.35	1.25	0.27	0.78	0.64	0.20	0.78
5	24	62	0.65	0.41	1.66	0.29	1.07	0.47	0.78	1.45	1.26	0.29	0.82	0.71	0.22	0.88
6	24	62	0.71	0.46	1.52	0.33	1.08	0.49	0.79	1.48	1.33	0.30	0.82	0.75	0.23	0.94
7	22–45	62	0.52	0.38	1.37	0.24	1.05	0.45	0.77	1.50	1.23	0.33	0.86	0.75	0.20	0.85
8	21	87	0.65	0.41	1.59	0.24	1.14	0.48	0.72	1.51	1.33	0.38	0.90	0.78	0.26	0.84
											Analyzed phytase, FTU/kg
Dataset No.	Approx. BW at sampling, kg	Limestone solubility at 5 min, %	Asp (total)	Cys (total)	Glu (total)	Gly (total)	Ser (total)	Tyr (total)	Pro (total)	Ala (total)	NC	NC +250	NC +500	NC +1,000	NC +2,000	NC +4,000
1	28	53	1.88	0.50	3.44	0.79	0.99	0.65	1.25	0.90	28	201	647	1,058	n.a.	n.a.
2	28	70	1.72	n.a.	4.86	0.92	1.01	0.51	1.50	0.89	347	617	906	1,494	2,289	n.a.
3	28	92	1.72	n.a.	4.93	0.92	1.02	0.52	1.53	0.90	326	682	1,001	1,417	1,958	n.a.
4	24	62	1.52	0.25	2.79	0.65	0.63	0.53	0.95	0.78	242	n.a.	806	1,370	2,379	4,347
5	24	62	1.59	0.33	3.10	0.78	0.68	0.54	1.07	0.86	188	n.a.	696	1,309	2,406	4,680
6	24	62	1.50	0.38	3.22	0.85	0.71	0.55	1.16	0.90	183	n.a.	682	1,220	2,240	4,239
7	22–45	62	1.62	0.29	3.08	0.73	0.73	0.57	1.04	0.89	147	432	564	1,141	1,810	4,850
8	21	87	1.84	0.24	3.17	0.71	0.88	0.60	1.09	0.95	135	440	1,429	2,535	n.a.	n.a.

n.a. not analyzed.

Within all datasets, analyzed dietary P was close to formulated values, but analyzed Ca was more variable. The in vitro solubility of the limestones used in the experiments ranged from 53 to 92% at 5 min (Kim et al, 2019; Table 1).

Modeling the percentage AID of AA against increasing phytase dose

A total of 18 individual AA was analyzed in all experiments, including Arg, His, Ile, Leu, Lys, Met, Phe, Thr, Trp, Val, Ala, Asp, Cys, Glu, Gly, Ser, Pro, and Tyr. The AID of total AA in the NC diet without added phytase ranged from 57.5% in dataset 1 to 86.3% in dataset 3 (data not shown).

The percentage AID of individual and total AA was modeled against increasing analyzed PhyG dose, by log-linear regressions fitted to the combined data from all 8 datasets. The results are presented in Table 3. For the NC diet without added phytase, the model-predicted AID of total and individual AA varied from a minimum of 36.2% for Gly to a maximum of 81.6% for Met. The mean AID of total AA was 68.6%. The AID percentage of total AA and of Arg, Asp, Cys, Gly, His, Ile, Lys, Phe, Ser, Thr, Trp, and Tyr all increased log-linearly with increasing PhyG within the dose range 0 to 4,000 FTU/kg (P < 0.05), whereas the AID percentage of Ala, Glu, and Leu tended to increase log-linearly (P = 0.089, P = 0.058 and P = 0.052, respectively; Table 3). By this analysis, at 1,000 FTU/kg, the model predicted that PhyG increased the AID of total AA by 3.8 percentage units compared with the NC (from 68.6 to 72.4%), and at 4,000 FTU/kg the predicted increase was 4.5 percentage units (from 68.6 to 73.1%) vs. no PhyG.

Table 3.

Estimated apparent ileal amino acid digestibility (%) in response to increasing analyzed phytase activity¹, modeled for all eight datasets by log-linear regression; results of the first statistical modeling analysis

	Phytase (PhyG) level, FTU/kg2						Log Linear P- value
	0	250	500	1,000	2,000	4,000
Amino acid
Ala	65.30	67.78	68.09	68.40	68.71	69.02	0.089
Arg	81.74	83.95	84.23	84.51	84.79	85.06	0.007
Asp	68.88	71.61	71.95	72.29	72.64	72.98	0.010
Cys	43.34	48.55	49.20	49.86	50.51	51.17	0.042
Glu	76.54	78.47	78.72	78.96	79.20	79.45	0.058
Gly	36.21	44.72	45.78	46.85	47.92	48.99	0.002
His	72.38	74.81	75.12	75.42	75.73	76.03	0.043
Ile	70.58	73.33	73.68	74.02	74.37	74.72	0.036
Leu	71.20	73.74	74.05	74.37	74.69	75.01	0.052
Lys3	78.78	80.67	80.91	81.15	81.39	81.62	0.024
Met3	81.64	82.77	82.91	83.05	83.20	83.34	0.254
Phe	72.37	75.18	75.53	75.88	76.23	76.59	0.011
Pro	65.93	66.84	66.95	67.07	67.18	67.30	0.583
Ser	65.56	69.07	69.51	69.95	70.39	70.83	0.022
Thr3	60.77	64.32	64.77	65.22	65.66	66.11	0.036
Trp	63.38	69.90	70.72	71.54	72.35	73.17	<0.001
Tyr	71.17	74.01	74.36	74.72	75.08	75.43	0.026
Val	64.86	67.34	67.65	67.96	68.28	68.59	0.127
Total AA	68.62	71.61	71.98	72.35	72.73	73.10	0.021

¹Above the analyzed activity in the negative control diet without added phytase.

²Phytase dose levels were log transformed prior to performing the regression and then back-transformed afterwards.

Relationship between AID of CP and phytate (IP_3-6); an example from Experiment 4

Figure 1 shows how the AID of CP (%) increased linearly (P < 0.05) with increasing phytate digestibility (AID IP_3-6) of the diet in Experiment 4 (the AID IP_3-6 was calculated at each analyzed phytase level and the data modeled against AID of CP).

Figure 1.

The relationship between apparent ileal digestibility (AID) of crude protein (%) and AID IP_3-6 (in diets with phytase dose levels between 0 and 4,000 FTU/kg), in pigs of ~24 kg BW fed corn-soybean meal-based diets containing 0.23% phytate-P (P < 0.001); data from Experiment 4.

Modeling the percentage unit improvement in AID of AA in response to increasing phytase level

Figure 2 shows the predicted percentage unit increase above the response to the NC diet (using data generated from the second modeling exercise) in the AID of A) indispensable, and B) dispensable AA, in response to increasing PhyG dose, as modeled by log-linear regression on the combined datasets. The associated parameters and statistical significance of the regressions are presented in Table 4. In this analysis, the estimated improvements for Lys, Met, and Thr excluded the contribution from the synthetic form of these AA. The fitted linear relationships were statistically significant (P < 0.05) for all individual AA and total AA. In all cases, increasing the dose of PhyG within the range of 250 to 4,000 FTU/kg resulted in increases in the improvements in AID of AA. Amongst the indispensable AA (Figure 2A), the improvements in AID with PhyG at 1,000 and 2,000 FTU/kg were greatest for Trp (+ 6.1 and + 6.8 percentage units, respectively) followed by Thr (+ 3.4 and + 3.8 percentage units, respectively) and least for His (+ 1.8 and + 2.0 percentage units, respectively) and Met (+ 1.8 and + 2.0 percentage units, respectively). Among the dispensable AA (Figure 2B), improvements in AID with PhyG at 1,000 and 2,000 FTU/kg were greatest for Gly (+ 7.5 and + 8.3 percentage units, respectively) and Cys (+ 5.6 and + 6.2 percentage units, respectively) and least for Glu (+ 2.2 and + 2.4 percentage units, respectively) and Ala (+ 2.5 and + 2.7 percentage units, respectively).

Figure 2.

Predicted percentage unit increase¹ above the response to the negative control diet in the apparent ileal digestibility of A) indispensable² and B) dispensable amino acids, in response to increasing phytase dose³, based on log-linear regression⁴; results of the second statistical modeling analysis. ¹Calculated individually for each of the eight datasets to account for between-dataset differences in AID of AA in the negative control diet without phytase. The resulting data were then pooled and modeled together via linear regression.²For Lys, Met and Thr the calculated percentage unit increase in AID excludes the concentration of supplemental synthetic AA. ³Above the analyzed activity in the negative control without phytase.⁴Phytase doses were log transformed prior to performing the regression and then back-transformed afterwards

Table 4.

Parameters of the fitted log-linear relationship¹ between the percentage point improvement (above the respective negative control) in apparent ileal digestibility of individual and total amino acids (Y) and increasing analyzed phytase activity (X)²; results of the second statistical modeling analysis

	Intercept	Slope	Log Linear P-value2
Item	a	b
Amino acid
Ala	0.0014	0.0083	0.022
Arg	-0.0022	0.0072	<0.001
Asp	-0.0015	0.0124	<0.001
Cys	0.0011	0.0185	<0.001
Glu	-0.0014	0.0078	0.008
Gly	-0.0025	0.0258	<0.001
His	-0.0012	0.0063	0.005
Ile	-0.0004	0.0112	<0.001
Leu	-0.0023	0.0096	<0.001
Lys3	0.0005	0.0099	<0.001
Met3	0.0001	0.0066	<0.001
Phe	-0.0020	0.0105	<0.001
Pro	0.0004	0.0115	0.002
Ser	0.0000	0.0132	<0.001
Thr3	-0.0011	0.0132	<0.001
Trp	-0.0026	0.0214	<0.001
Tyr	-0.0011	0.0120	<0.001
Val	-0.0016	0.0100	<0.001
Total AA	-0.0011	0.0084	0.001

¹The linear equation is: Y = b* logX + a.

²Phytase levels were log transformed prior to performing the regression and then back-transformed afterwards.

³The calculated AID percentage improvements for these AA excluded the concentration of supplemental synthetic AA.

Prediction of expected AA contributions from PhyG based on the modeled improvements in AID of AA

An example of how the modeled relationship between the percentage unit improvements in AID of AA and increasing phytase dose can be used to predict the expected AA contributions of the phytase when applied at a given dose to a U.S. type diet, an EU type diet or an Asian type diet, is given in Figure 3. According to this example, the predicted digestible Lys contribution (matrix value) of PhyG dosed at 1,000 FTU/kg in a U.S. type diet (corn-soybean meal) is 0.028% whereas that of Met + Cys is 0.024% and those of Thr and Trp are, respectively, 0.026% and 0.015%. The predicted digestible AA contributions (matrix values) would be impacted by total AA content as well as the proportion of synthetic AA that is to be included in the diet. This can be factored into the calculation, as shown in the example (Figure 3).

Figure 3.

Predicted digestible amino acid (AA) contributions of three different pig diet compositions when supplemented with 1,000 FTU/kg of PhyG phytase, calculated based on the total AA concentration and synthetic AA inclusion of the diet and the log linear model-predicted improvement in AID of AA (percentage unit improvement) by the phytase shown in Fig. 2. Main ingredients of the U.S. diet: corn, 55.4%; soybean meal, 17.8%; whey permeate, 15.0%; soybean protein concentrate, 5.0%; blood meal, 3.0%; limestone, 1.4%; monocalcium phosphate, 1.0%. Main ingredients of the EU diet: barley, 28.6%; corn, 27.0%; wheat, 15.0%; soybean meal, 20.0%; peas, 4.0%; fat, 1.9%: limestone, 0.9%; monocalcium phosphate, 0.5%. Main ingredients of the Asian diet: corn, 55.0%; soybean meal 25.0%; wheat, 10.0%; tallow, 3.8%, corn gluten meal, 3.0%; dicalcium phosphate, 1.3%; limestone, 0.4%. As an example: for calculation of the expected contribution of Lys: in a US-style diet containing 1.5% total Lys, with Lys HCL inclusion at 0.45%, the actual content of synthetic Lys would be 0.45 × 78% (the purity of Lys) = 0.35%. Therefore, the content of Lys excluding the synthetic component would be 1.50–0.35 = 1.15%. According to the fitted linear regression for the effect of PhyG on apparent ileal digestibility of Lys shown in Fig. 2A and Table 9, the expected Lys contribution by PhyG at 1,000 FTU/kg is 2.5% of 1.15 = 0.028% g/kg (or 0.28 g/kg).

Discussion

To the best of our knowledge, the current analysis represents the first published study in which AA digestibility responses to phytase in pigs, determined using data from multiple trials of the same phytase, have been modeled together. Such analyses form the basis of digestible AA matrix generation by phytase manufacturers and are important for generating a matrix that is both robust and yet can be tailored to the specific diet to which the phytase is to be added. The deliberate inclusion among the 8 datasets of variation in diet composition (cereal and oilseed composition, PP content, limestone solubility), diet form (mash or pellet), pig sex and breed, was aimed at being able to account for some of the major factors known to cause variation in animal responses to dietary interventions in the field and hence at deriving matrix values that would be more robust in production practice. A similar modeling approach was recently undertaken for the same phytase using data from multiple experiments conducted with broilers (Dersjant-Li et al., 2022b). Analyzed phytase levels were used to model the dose-response to improve the accuracy of model-derived predictions.

In the first analysis, AID percentages were modeled unadjusted for among-experiment differences in the response to the NC. The digestibility of AA by pigs and broilers in mixed diets is inherently linked to the AA composition and digestibility of the individual protein substrates present in a given diet and it is assumed that there is additivity among ingredients in mixed diets (Stein et al., 2005). The digestibility of AA varies among diets depending on the raw material composition of the diet, which also affects the dynamics of AA release and absorption by the animal (Chen et al., 2019; Lee et al., 2020; Norgaard et al., 2021; Hu et al., 2022). This is one of the main reasons why wide variation in AA digestibility values is obtained from different experiments, as illustrated by the systematic review of Cowieson et al. (2017) in which individual AA digestibility coefficients in the control (unsupplemented) diets from 28 experiments ranged from ~0.35 to ~0.95, with most datapoints clustered between 0.6 and 0.9, which is still a wide range. The average digestibility coefficient for the sum of all AA across all 28 experiments reported by Cowieson et al. was 0.75 ± 0.013. In the present study, the response to the NC diet also varied but to a lesser extent; the AID of total AA ranged from 57.5% (in dataset 1) to 86.3% (in dataset 2; data not shown).

The results of the first modeling analysis showed a clear log-linear improvement in the AID of total AA and of 12 of 18 individual AA with PhyG supplemented within the range from 250 to 4,000 FTU/kg, indicating a beneficial effect of the phytase on AA digestibility. This is consistent with individual reports of the effect of PhyG and certain other phytases on AA digestibility in pigs in which improvements in the AID of most individual AA and total AA have been observed (Espinosa et al., 2022; Lagos et al., 2022), but is in contrast with the results of other studies where no beneficial effect on AA digestibility was observed (She et al., 2018; Arredondo et al., 2019; Mesina et al., 2019). Such differences may be due to differences in the properties and main functional area of different phytases because different phytases exhibit differing efficacy when applied at an equivalent dose (Dersjant-Li and Kwakernaak, 2019; Dersjant-Li et al., 2021). The PhyG phytase has a low pH optimum (pH 3.5–4.5 in vitro) and wide pH range of activity (pH 2.0–5.0; Christensen et al., 2020) that make it well suited to the low pH conditions of the pig stomach (~pH 3.0; Li et al., 2008) where phytate hydrolysis by phytase predominantly occurs. This phytase also has a high P-release capacity in pigs (estimated as 1.83 grams of digestible P equivalence from monocalcium phosphate per kilogram of diet, on average, based on bone ash, weight gain, or feed conversion ratio when the phytase is dosed at 1,000 FTU/kg; Dersjant-Li et al., 2020). The model-predicted increases in AID of total AA with the phytase at 500 or 1,000 FTU/kg from the first modeling analysis compare favorably with the observed improvements in total AA digestibility coefficients reported for the 28 experiments analyzed by Cowieson et al. (2017), in which doses ranged from 250 FTU/kg to 20,000 FTU/kg although most observations were at 500 or 1,000 FTU/kg. Reasons for the apparently greater improvement in AID of total AA with PhyG compared with the average response reported in the Cowieson et al. (2017) systematic review may include that the present analysis focused on multiple studies of a single phytase rather than of multiple phytases, the latter of which is likely to have introduced greater variation in the results.

No significant relationship between dietary phytate-P level and the percentage point increase in AID of AA (individual or total) was observed in the present analysis. Hence, dietary phytate-P content was not included as a factor in the final model. Such a relationship was also not observed in a comparable meta-analysis of the effect of PhyG on the AID of AA in broilers, although a clear relationship between phytate-P level and the increase in P digestibility from phytase has been shown in both species (Espinosa et al., 2021; Dersjant-Li et al., 2022b). Further research is needed to evaluate more fully the relationship between phytate-P level in the diet and the AA-‘release’ afforded by phytase. However, in practice, diets containing higher phytate-P levels are usually formulated with a higher total AA content to ensure the digestible AA requirement is met. As the generated digestible AA matrix uses the total AA content of the diet as a known input variable, in this way the matrix is indirectly adjusted for the dietary phytate-P level.

The observation that the AID of AA continued to increase incrementally up to the maximum dose (4,000 FTU/kg) for all individual AA does not agree with the conclusions by Cowieson et al. (2017) who observed no significant dose-response relationship between phytase supplementation and the coefficient of AID of total AA within the range 250 to 2,000 FTU/kg. In the present study, there was a log-linear improvement in AID of total AA with PhyG at 4,000 vs. 250 FTU/kg. The phytase used in this experiment enables a rapid and extensive breakdown of IP_3-6 (Christensen et al., 2020; Dersjant-Li 2022a). It is therefore hypothesized that the positive dose-response relationship between phytase supplementation and AID of total AA is a result of the phytase effecting incrementally greater increases in IP₆ degradation and that this reduced the availability of IP₆ and lower IP-esters to exert antinutritive effects on protein and AA availability and absorption. The observation from experiment 4 that there was a clear, positive, association between increasing digestibility of phytate (at increasing PhyG dose level) and improved AID of AA supports that the improved AA digestibility may have resulted from increased phytate degradation by PhyG.

The second modeling analysis revealed that phytase supplementation resulted in statistically significant improvements in AID above the NC diet for all individual AA and for total AA. The model-predicted improvements were greatest for those individual AA whose digestibility was low in the basal diet without phytase (Trp and Thr among the indispensable AA and Gly and Cys among the dispensable AA, although Ser, Tyr, and Pro were also highly improved) and lowest for Met whose digestibility was high in the basal diet. The observation that AA with low digestibility in the NC diet had the greatest improvement when phytase was added to the diet may reflect that there was greater capacity for an improvement by phytase in those AA whose inherent digestibility is low. This phenomenon was also observed in the comparable meta-analysis of broiler experiments performed by Dersjant-Li et al. (2022b). Also similar to the present study, the meta-analysis of pig studies by Cowieson et al. (2017) reported Trp and Thr to be among the AA whose digestibility was most improved by phytase and Met to be among the least improved by phytase. A meta-analysis by Zouaoui et al. (2018) that included 137 experiments with pigs equally observed a positive linear effect of phytase supplementation on ileal digestibility of Met and Thr. Improvements in Thr and Trp digestibility are relevant because these are among the most limiting AA in diets for pigs. Tryptophan is important in the development of weaned piglets, specifically in growth performance, intestinal mucosal barrier function, immune regulation, appetite and the stress response (Le Floc’h and Seve, 2007), whereas Thr is a major constituent of mucin secretions (Faure et al., 2005). Phytate in pig diets increases mucin secretions in the small intestine (Woyengo and Nyachoti, 2013), whilst evidence from broiler studies indicates that this may be reduced by phytase (Cowieson et al., 2004). Other studies using different phytases have found no such effect (Mesina et al., 2019), which may be due to the different pH optima of different phytases. The present data indicate that PhyG may have a relatively greater effect on those AA that are involved in endogenous protein losses via mucin secretion (Thr, Ser, and Pro) and less effect on Met, which is not a major component of mucin. Further studies are needed to confirm this and its relation to mucin production. The digestibility of Cys was greatly improved by PhyG whereas no effect of exogenous phytase on the ileal digestibility of Cys was reported by Cowieson et al. (2017). Whether this effect on Cys is unique to the PhyG phytase or was an anomaly in the data is unknown. The low digestibility of Cys in the NC diet compared with other AA may be the reason why a greater improvement was observed for Cys than for other AA. However, this was also evident in the control diets reviewed in Cowieson et al. (2017).

The modeled relationship between the percentage unit improvement in AID of AA and increasing phytase level enabled us to predict (as an example) the expected AA contribution of the phytase when applied at 1,000 FTU/kg to a US-, EU- or Asian-type of diet of known AA content. This calculated contribution is the amount of AA by which, in practice, the diet formulation could be reduced to account for the activity of the enzyme. The lack of a relationship between dietary PP level and the increase in AID of AA from phytase within the combined dataset means that the generated matrix values should be equally applicable to diets of differing PP levels. Their application to diets of higher digestible P content than that used here is less certain but there is no a priori reason to expect that a higher digestible P content in the basal diet would affect changes in the AID of AA in response to phytase supplementation. Although increasing dietary P content increases feed intake in pigs (Saraiva et al., 2009) and this may be expected to increase AA intake; increased AA digestibility in response to increased dietary available P content is not consistently shown in the literature and in broilers, increasing dietary P content has very limited impact on AA digestibility (Dersjant-Li and Kwakernaak, 2019). This may be because increasing P in the diet alone does not reduce phytate-AA interactions, therefore has limited effect on AA digestibility, whereas phytase reduces phytate-AA interactions resulting in the potential to increase AA digestibility. On this basis, it is expected that the digestible AA matrix values generated here would apply equally to diets with higher P content as to the P-deficient diet employed here.

It is considered that this approach to matrix derivation, which takes account of the actual feed formulation (and its associated AA concentration), will generate predicted digestible AA contributions that are more accurate than ‘one matrix fits all’ approach. The approach may also facilitate formulating diets with lower protein, thereby reducing N excretion from pig production. Finally, the approach may enable greater inclusion of alternative ingredient sources of protein and AA within the diet, by using the phytase to increase the AID of AA from those ingredients by a known amount. The next step is to validate the matrix values generated in this modeling exercise in growth performance experiments under commercial production settings.

In conclusion, the data obtained from this work support the use of PhyG phytase in swine diets for the improvement of AA digestibility. Modeling of the combined data from 8 datasets from 5 experiments that incorporated variation in diet composition, diet form, pig breed, age and sex revealed log-linear improvements in the AID of 18 individual AA and total AA when PhyG was supplemented in the range from 250 to 4,000 FTU/kg. Improvements were greatest for Trp, Thr, Gly, and Cys. An example of how the generated model can be used to make predictions of the digestible AA contribution from the phytase (matrix value) at a given dose when supplemented to different diets of known AA concentration is provided. Using the known AA concentration of the specific diet to which the phytase will be added as one of the input parameters allows for greater accuracy in matrix prediction and supports the derivation of diet-specific recommendations in commercial feed formulations.

Glossary

Abbreviations

AA

amino acids

AID

apparent ileal digestibility

ANF

anti-nutritional factors

ATTD

apparent total tract digestibility

BW

body weight

FTU

phytase units

HPLC

high performance liquid chromatography

ICP-MS

inductively coupled plasma mass spectrometry

ICP-OES

inductively coupled plasma mass spectrometry

NE

net energy

NC

negative control

ME

metabolizable energy

PP

phytate-phosphorus

PhyG

a consensus bacterial 6-phytase variant

SID

standardized ileal digestibility

STTD

standardized total tract digestibility

Conflict of interest statement. The authors declare no conflict of interest.

Author Contributions

Deepak Ettungalpadi Velayudhan (Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Writing - original draft, Writing - review & editing), Yueming Dersjant-Li (Conceptualization, Investigation, Methodology, Project administration, Writing - review & editing), Rosil Lizardo (Resources, Supervision, Visualization, Writing - review & editing), Arun Kumar (Resources, Supervision, Visualization, Writing - review & editing), Hans Stein (Resources, Supervision, Visualization, Writing - review & editing), Charmaine Espinosa (Visualization, Writing - review & editing), Vinicius Cantarelli (Resources, Supervision, Visualization, Writing - review & editing), Rachael Hardy (Writing - review & editing), Ester Vinyeta (Supervision, Writing - review & editing), and Leon Marchal (Supervision, Writing - review & editing)