CNCPS-Based Nutrient Fractionation, Ruminal Degradation Kinetics, and Intestinal Amino Acid Flow Estimates in Dairy Feedstuffs

Yansong Ge; Ziyao Wang; Xu Sun; Xueyan Lin; Zhonghua Wang

doi:10.3390/ani16152371

. 2026 Aug 3;16(15):2371. doi: 10.3390/ani16152371

CNCPS-Based Nutrient Fractionation, Ruminal Degradation Kinetics, and Intestinal Amino Acid Flow Estimates in Dairy Feedstuffs

Yansong Ge ^1,^†, Ziyao Wang ^1,^†, Xu Sun ¹, Xueyan Lin ^1,^*, Zhonghua Wang ^1,^*

Editor: Jian Ma¹

PMCID: PMC13465355 PMID: 42589011

Simple Summary

Nineteen dairy feedstuffs were characterized for CNCPS carbohydrate and protein fractions. Ruminal disappearance was measured in situ, and fraction-specific degradation constants were fitted. Exploratory regressions related feed composition to protein-fraction Kd values, which were then used, together with in situ-fitted Kd values, to generate two sets of CNCPS-based intestinally absorbable amino acid flow estimates. These estimates were model-derived rather than directly measured. Because the regressions were developed from only 19 feedstuffs and lacked external validation, the equations should be regarded as exploratory relationships, not practical prediction tools.

Keywords: CNCPS, ruminal degradation, protein fractions, feed evaluation, metabolizable protein, intestinal amino acid flows, small intestine, dairy cows

Abstract

To improve feed evaluation for precision dairy nutrition, this study integrated chemical analysis, CNCPS fractionation, in situ ruminal disappearance, exploratory regression, and amino acid-flow calculations for 19 dairy feedstuffs. Conventional nutrients and CNCPS fractions were determined, and total CP and carbohydrate-fraction disappearance were measured in four ruminally cannulated Holstein cows. Protein-fraction Kd values were inferred by nonlinear fitting with measured CNCPS fractions held constant. In the three silages, non-protein nitrogen represented 93.82–96.80% of soluble CP. Energy feeds had greater total carbohydrate and starch-associated CB1 fractions, whereas cottonseed had the greatest unavailable cell-wall fraction. Exploratory equations for Kd1, Kd2, and Kd3 yielded in-sample R² values of 0.7422, 0.7365, and 0.3871, respectively. Two CNCPS-based sets of intestinally absorbable essential amino acid estimates were then calculated using either in situ-fitted or regression-estimated Kd values. The regression-assisted estimates were generally comparable with those based on in situ-fitted Kd values. However, neither set represented direct intestinal measurements, agreement between them did not constitute validation, and Kp was derived from NRC equations under a standardized dietary scenario. Thus, the amino acid estimates are scenario-dependent, and the regression equations should be considered exploratory pending external validation.

1. Introduction

Accurate feed evaluation is essential in precision dairy nutrition because metabolizable protein and intestinally absorbable essential amino acids influence milk protein synthesis and nitrogen-use efficiency [1,2]. Adequate amino acid supply also contributes to reducing nitrogen excretion [3]. Reliable estimates of ruminal protein degradation and intestinal amino acid supply are therefore required to compare feed protein value, match nutrient supply with animal requirements, and avoid unnecessary dietary CP [4,5].

The CNCPS partitions feed carbohydrate and protein into fractions differing in solubility and ruminal degradation [6], but its outputs depend on feed composition, fraction data, and degradation parameters [7]. Estimates may also vary among feed evaluation systems [8], while metabolizable protein supply and dietary composition can affect nutrient utilization and production responses [9,10]. Existing CNCPS libraries and NRC/NASEM tables provide useful reference values, yet integrated data generated from the same commercial feed samples remain limited, particularly for Chinese feeds and processed by-products. Previous studies have generally focused on individual forages or feed ingredients [11,12]. An integrated in situ and CNCPS approach has also been applied to an individual grain source [13]. However, contemporaneous measurements of chemical composition, CNCPS fractions, fraction-specific in situ Kd values, insoluble-protein amino acid composition, and model-derived intestinal amino acid flows across a common multi-feed dataset remain scarce.

Accordingly, this study evaluated 19 dairy feedstuffs using standardized sampling, analytical, and in situ procedures. The descriptive objectives were to characterize CNCPS carbohydrate and protein fractions and determine fraction-specific ruminal degradation kinetics. The modeling objectives were to explore associations between feed characteristics and fitted protein-fraction Kd values and to calculate scenario-dependent intestinally absorbable essential amino acid flows within the CNCPS framework. The study was intended to provide feedstuff-level integration of batch-derived composition data, representative-composite in situ measurements, amino acid composition, and model outputs, rather than to replace existing feed libraries or establish a universally applicable prediction system.

2. Materials and Methods

2.1. Evaluation of Carbohydrate and Protein Fractions in Common Ruminant Feedstuffs

2.1.1. Experimental Materials and Feed Sampling

Nineteen feedstuffs commonly used in dairy cow diets were collected or purchased between late May and December 2024 from commercial dairy farms, feed suppliers, or feed-processing enterprises in Shandong, Beijing, Tianjin, Shanxi, and Henan, China. Each feedstuff was represented by at least three independent production or commercial batches. Five spatially separated subsamples from each batch were combined into one composite sample; batch composites were retained and analyzed separately.

The feedstuffs were soybean meal, wheat, flaked corn, wheat bran, extruded soybean hulls, molassed soybean hulls, pelleted sugar beet pulp, dried distiller grains with solubles (DDGS), wheat middlings, cottonseed 1, cottonseed 2, whole-plant corn silage 1, whole-plant corn silage 2, corn stover silage, wheat straw, alfalfa, dried brewer grains, dried distiller grains 1, and dried distiller grains 2. Products sharing a general name but bearing different numerical designations originated from independent sources or processing batches and were treated as distinct commercial feed products rather than biological replicates. Specifically, the two whole-plant corn silages originated from separate commercial sources and ensiling units, the two cottonseed samples from different regions and suppliers, and the two dried distiller grains from different manufacturers and batches; the latter were also distinct from DDGS. Detailed information on sample sources, collection periods, feed types, storage conditions, and processing characteristics is provided in Table A1.

Dry samples were sealed in polyethylene bags and stored at room temperature in a cool, dry, and dark environment. Silages were sealed immediately after collection and stored at −20 °C. Original silage subsamples were used to determine as-fed DM, and the remaining material was freeze-dried, ground, and used for chemical analyses.

2.1.2. Determination Parameters

The determined parameters included dry matter (DM), crude protein (CP), ether extract (EE), ash, neutral detergent fiber (NDF), acid detergent fiber (ADF), lignin, starch, non-protein nitrogen (NPN), soluble crude protein (SCP), neutral detergent insoluble protein (NDIP), and acid detergent insoluble protein (ADIP). All conventional nutrient and CNCPS fraction analyses were performed in triplicate for each independent batch composite.

2.1.3. Major Reagents

Unless otherwise stated, all chemicals and reagents were of analytical grade and were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). These included concentrated sulfuric acid, hydrochloric acid, sodium hydroxide, disodium hydrogen phosphate, sodium dihydrogen phosphate, trichloroacetic acid, perchloric acid, sodium dodecyl sulfate, disodium EDTA, sodium tetraborate, cetyltrimethylammonium bromide, ethylene glycol monoethyl ether, acetone, diethyl ether, and ethanol.

2.1.4. Main Equipment

Major equipment included an analytical balance with a readability of 0.0001 g (Mettler Toledo, Greifensee, Switzerland), a muffle furnace (Nabertherm GmbH, Lilienthal, Germany), a plant sample grinder (Retsch GmbH, Haan, Germany), a fiber digestion apparatus (ANKOM Technology, Macedon, NY, USA), a Soxhlet extraction system and an automatic Kjeldahl nitrogen analyzer (VELP Scientifica Srl, Usmate Velate, Italy), and a UV–visible spectrophotometer (Shimadzu Corporation, Kyoto, Japan). Feed samples used for chemical analyses were ground to pass through a 0.45 mm sieve.

2.1.5. Analytical Methods for Nutrients and CNCPS Fractions

Determination of Conventional Nutrients

The DM contents of feed samples and ruminal residues were determined by oven drying to constant weight according to GB/T 6435-2014 [14]. CP was measured by the Kjeldahl method according to GB/T 6432-2018 [15], using a nitrogen-to-protein conversion factor of 6.25. NDF and ADF were determined using neutral- and acid-detergent extraction procedures according to GB/T 20806-2022 [16] and NY/T 1459-2022 [17], respectively. Crude ash was determined by combustion in a muffle furnace according to GB/T 6438-2007 [18], and EE was measured by solvent extraction according to GB/T 6433-2006 [19]. A summary of the analytical procedures and corresponding methodological standards is provided in Table A6. Acid detergent lignin (ADL) content was determined according to the procedure described by van Soest et al. [20].

Determination and Calculation of CNCPS Fractions

Total carbohydrate (CHO) content was calculated on a DM basis as follows:

CHO (% DM) = 100 − CP − EE − Ash

The CNCPS carbohydrate fractions were calculated as follows:

CB1 (% CHO) = Starch (% DM)/CHO (% DM) × 100

CC (% CHO) = [2.4 × ADL (% DM)]/CHO (% DM) × 100

CB2 (% CHO) = {[NDF (% DM) − NDIP (% DM) − 2.4 × ADL (% DM)]/CHO (% DM)} × 100

CA (% CHO) = 100 − CB1 − CB2 − CC

In the present calculations, CB1 was calculated from the measured starch concentration, CC represented the unavailable cell-wall carbohydrate fraction, CB2 represented the potentially available cell-wall carbohydrate fraction, and CA was obtained by difference. The combined non-structural carbohydrate fraction was calculated as:

CNSC (% CHO) = CA + CB1

Starch content was determined using the perchloric acid–anthrone colorimetric method. The CNCPS protein fractions were calculated as follows:

PA (% CP) = NPN (% SCP) × SCP (% CP)/100

PB1 (% CP) = SCP (% CP) − PA (% CP)

Before calculating PB2, PB3, and PC, NDIP and ADIP were converted from a DM basis to a CP basis as follows:

NDIP (% CP) = NDIP (% DM)/CP (% DM) × 100

ADIP (% CP) = ADIP (% DM)/CP (% DM) × 100

The remaining protein fractions were then calculated as:

PC (% CP) = ADIP (% CP)

PB3 (% CP) = NDIP (% CP) − ADIP (% CP)

PB2 (% CP) = 100 − PA − PB1 − PB3 − PC

PA represents the non-protein nitrogen fraction, PB1 the rapidly degradable true protein fraction, PB2 the intermediately degradable protein fraction, PB3 the slowly degradable fiber-bound protein fraction, and PC the unavailable acid-detergent insoluble protein fraction. Non-protein nitrogen, protein soluble in borate–phosphate buffer, neutral detergent insoluble protein, and acid detergent insoluble protein were determined according to van Soest et al. [20].

For both conventional nutrient and CNCPS fraction analyses, each independent batch composite was analyzed in triplicate. The three analytical replicates were averaged to obtain one batch-level value. For each feedstuff, results are presented as the mean ± standard deviation of the independent batch-level values, and analytical replicates were not treated as independent observations.

2.2. Evaluation of Ruminal Degradation Rate Constants of Feed Carbohydrate and Protein Fractions

2.2.1. Experimental Materials

The same feedstuffs described in Section 2.1.1 were used. For each feedstuff, the independent batch composites used for chemical composition and CNCPS fraction analyses were combined in equal proportions on a DM basis to prepare one representative feedstuff-level composite sample. This representative composite was used for the subsequent in situ ruminal incubation experiment.

2.2.2. Experimental Animals and Management

Four multiparous, mid-lactation Holstein dairy cows fitted with permanent ruminal cannulas were used in this experiment. The cows had similar physiological status, with body weight, days in milk, milk yield, and dry matter intake of 514 ± 74 kg, 147.46 ± 28.36 d, 42.61 ± 6.80 kg/d, and 26.39 ± 4.20 kg/d, respectively. The cows were clinically healthy and were housed and managed under the same feeding conditions throughout the experimental period. The cows were fed a lactation-period total mixed ration twice daily at 08:30 and 15:30 and milked twice daily at 06:30 and 18:30. The forage-to-concentrate ratio of the basal diet was 41.62:58.38 on a DM basis, and the ingredient composition and nutrient levels of the basal diet are presented in Table A2. All cows had free access to water. Ruminal pH was measured during the experimental period using a calibrated portable pH meter (Mettler Toledo, Greifensee, Switzerland). The ruminal pH ranged from 6.2 to 6.8, which was within the normal physiological range for dairy cows. The ruminally cannulated cows showed normal feed intake, rumination, and health status throughout the experiment.

2.2.3. In Situ Nylon Bag Procedure

Ruminal degradation was determined using the in situ nylon bag technique. Representative composites were ground through a 2.0 mm screen to standardize preparation and limit particle loss through the 50 μm bag pores. Concentrate (5 g) or roughage (3 g) samples were weighed into 8 cm × 12 cm nylon bags. Bags containing the same feedstuff were placed in approximately 50 cm nylon mesh carriers, positioned in the ventral rumen, and secured to the cannula.

Nineteen sequential runs were conducted, one feedstuff per run. In each run, all bags were inserted simultaneously into the four cows before morning feeding and removed sequentially at the assigned times. One bag was incubated per cow at each time point; thus, the four cows served as biological replicates for the disappearance measurements, and the bags were measurement units within cows. The final Kd parameters were estimated at the feedstuff level from the mean disappearance series across the four cows, as described in Section 2.2.6. Incubation times were 1, 2, 4, 8, 16, 24, 36, and 48 h for concentrates and 2, 4, 8, 16, 24, 36, 48, and 72 h for roughages; the additional early and late time points captured rapid concentrate and slow roughage disappearance, respectively.

Four zero-hour bags per feedstuff were immediately washed, dried, and weighed without ruminal incubation to quantify procedural washing loss and immediate particle loss. Each run also included four blank bags, one per cow, incubated for the longest assigned time and processed identically to monitor bag-weight changes and background material. Because blank-bag weight changes were negligible, no numerical correction was applied.

Because particle structures differed among feedstuffs, all bags were washed by the same trained operator in five successive 1 min cycles with fresh clean water until the rinse water was clear. Bags were then dried at 65 °C to constant weight for approximately 72 h, weighed, and analyzed as described in Section 2.1.5. Microbial N contamination was corrected using ¹⁵N-labeled ruminal solid-associated bacteria (SAB). Ruminal microorganisms were initially labeled by continuous intraruminal infusion of ¹⁵N-labeled ammonium sulfate (99 atom% excess; Cambridge Isotope Laboratories, Inc., Tewksbury, MA, USA) at 3 g/d for 6 d before the first incubation run [21]. Thereafter, the infusion was maintained at the same rate throughout all 19 sequential incubation runs to maintain microbial ¹⁵N enrichment. In each run, SAB was collected separately from each cow, isolated from solid digesta as described by Legay-Carmier and Bauchart [22], and matched to residues from the same cow. The ¹⁵N atom-percent excess of residues and SAB was determined using an elemental analyzer (Elementar Analysensysteme GmbH, Langenselbold, Germany) coupled to a DELTA V Advantage isotope-ratio mass spectrometer (Thermo Fisher Scientific, Bremen, Germany). Microbial N was estimated from the ratio of ¹⁵N atom-percent excess in residual N to that in SAB N and subtracted from total residual N before calculating corrected CP disappearance and degradation kinetics [23].

2.2.4. Experimental Instruments and Reagents

The equipment and reagents used in this experiment were the same as those described in Section 2.1.3 and Section 2.1.4, with the addition of a 2.0 mm screen, ¹⁵N-labeled ammonium sulfate, an intraruminal infusion apparatus, and the elemental analyzer–isotope-ratio mass spectrometry system described in Section 2.2.3.

2.2.5. Analytical Methods

The determined parameters and analytical methods were the same as those described in Section 2.1.2 and Section 2.1.5.

2.2.6. Calculation Methods

The proportion of microbial N in total residual N was calculated as follows:

P_MN = ¹⁵N′_residue/¹⁵N′_SAB

The amount of microbial N associated with each incubation residue was calculated as follows:

N_microbial = N_residue × P_MN

Feed-derived residual N was calculated as follows:

N_feed,residue = N_residue − N_microbial

The N disappearance rate corrected for microbial contamination was calculated as follows:

tDR(%) = [(N_initial − N_residue + N _microbial)/N_initial] × 100

where P_MN is the proportion of microbial N in total residual N; ¹⁵N′_residue and ¹⁵N′_SAB are the ¹⁵N atom-percent excess values of residual N and SAB N, respectively; N_initial is the initial amount of N placed in the nylon bag; N_residue is the total amount of residual N after incubation; N_microbial is the estimated amount of microbial N associated with the residue; and N_feed,residue is the corrected amount of feed-derived residual N. The corrected N disappearance values were used to calculate corrected CP disappearance and protein degradation kinetics.

The ruminal degradation kinetics were described using the model proposed by Ørskov and McDonald [24]:

P = a + b (1 − e^−ct)

where P is the degradability at time t, a is the rapidly soluble fraction, b is the slowly degradable fraction, and c is the degradation rate constant of fraction b.

The zero-hour-bag values were used only to quantify procedural washing loss and were not used to define or replace the chemically determined PA fraction. Blank nylon bags were used to monitor changes in bag mass and background material during incubation, washing, and drying. Because these changes were negligible, no numerical blank-bag correction was applied. Residual N values were corrected for microbial contamination before calculating CP disappearance and fitting the protein degradation curves.

Within the CNCPS framework, crude protein is divided into five fractions: PA, PB1, PB2, PB3, and PC. In this equation, A, B1, B2, and B3 represent the proportions of PA, PB1, PB2, and PB3 in total crude protein, respectively, and are expressed as % CP rather than absolute nutrient amounts. Accordingly, A was fixed exclusively at the experimentally determined PA proportion and was not derived from the zero-hour washing loss. Because fraction PC is considered undegradable in the rumen, the ruminal degradation equation for crude protein was expressed as follows:

P = A + B1 (1 − e^−Kd1t) + B2 (1 − e^−Kd2t) + B3 (1 − e^−Kd3t)

For each feedstuff and incubation time, the corrected total CP disappearance values obtained from the four cows were averaged to generate one feedstuff-level mean disappearance value. The resulting mean disappearance series, Pobs(t), was used as the dependent dataset for a single nonlinear fit for each feedstuff. The experimentally determined proportions of PA, PB1, PB2, and PB3 in total CP were fixed and entered into the model as A, B1, B2, and B3, respectively, whereas PC was excluded because it was considered ruminally undegradable. Kd1, Kd2, and Kd3 were estimated simultaneously by nonlinear least-squares fitting in GraphPad Prism version 5.0 for Windows (GraphPad Software Inc., La Jolla, CA, USA) by minimizing the sum of squared differences between the observed and model-predicted mean CP disappearance values across all incubation times. Thus, the individual protein fractions were not measured separately, and the reported Kd values represent feedstuff-level fitted constants derived from the mean disappearance curves of the four cows rather than means of cow-specific Kd estimates.

For example, soybean meal contained 7.81%, 6.78%, 83.01%, and 0.46% of CP as PA, PB1, PB2, and PB3, respectively. Its fitted equation was therefore:

Ppred(t) = 7.81 + 6.78 (1 − e^−Kd1t) + 83.01 (1 − e^−Kd2t) + 0.46(1 − e^−Kd3t).

The three Kd values were adjusted jointly until the predicted total CP disappearance curve provided the best least-squares fit to the corrected CP disappearance values observed over the complete incubation period. The PC fraction, which accounted for 1.94% of CP in soybean meal, was not included because it was assumed to be undegradable in the rumen.

For carbohydrates, the CNCPS classifies feed carbohydrates into four fractions: CA, CB1, CB2, and CC. Fraction CC is considered unavailable and was therefore not assigned a ruminal degradation rate constant. For each feedstuff and incubation time, the concentrations of CA, CB1, and CB2 in the original feed and incubation residue were calculated using the CNCPS fractionation equations described in Section Determination and Calculation of CNCPS Fractions and were expressed on a DM basis.

The initial amount of each carbohydrate fraction placed in the nylon bag was calculated as follows:

M_initial,i = W_initial × (C_initial,i/100)

The amount of the corresponding fraction remaining after incubation time t was calculated as follows:

M_residual,i,t = W_residual,t × (C_residual,i,t/100)

The disappearance of each carbohydrate fraction at incubation time t was then calculated as follows:

D_i,t (%) = [(M_initial,i − M_residual,i,t)/M_initial,i] × 100

where M_initial,i is the initial amount of carbohydrate fraction i placed in the nylon bag; M_residual,i,t is the amount of fraction i remaining after incubation time t; W_initial is the initial DM weight of feed placed in the nylon bag; W_residual,t is the residual DM weight after incubation, washing, and drying; and C_initial,i and C_residual,i,t are the concentrations of fraction i in the original feed and incubation residue, respectively, expressed as % DM. Thus, disappearance was calculated from the initial and residual amounts of each fraction rather than from changes in concentration alone.

Analytical variation in residual DM weight or chemical composition could theoretically result in negative disappearance values or values greater than 100%, particularly when the initial amount of a fraction was small. Therefore, all calculated disappearance values were checked against the original bag weights, residual weights, analytical results, and calculation records before nonlinear curve fitting. Suspect values were rechecked using the original laboratory records, and no values were automatically truncated, replaced, or imputed. For each feedstuff and incubation time, the verified carbohydrate-fraction disappearance values obtained from the four cows were averaged. The resulting feedstuff-level mean disappearance series were fitted by nonlinear least-squares regression to estimate Kd4, Kd5, and Kd6 for CA, CB1, and CB2, respectively. Accordingly, the reported carbohydrate-fraction Kd values are feedstuff-level fitted constants rather than means of cow-specific estimates.

2.3. Estimation of Intestinally Absorbable Amino Acid Flows Based on the CNCPS Model

2.3.1. Feed Samples for Amino Acid Analysis

The same feedstuffs described in Section 2.1.1 were used.

2.3.2. Amino Acid Analysis Instruments

Amino acid analysis was performed using an L-8800 automatic amino acid analyzer (Hitachi High-Tech Corporation, Tokyo, Japan). The analytical balance and sample-preparation equipment were the same as those described in Section 2.1.4.

2.3.3. Calculation of Intestinally Absorbable Amino Acid Flows

Calculation Method of Rumen Passage Rate

The passage-rate equations used in this study were derived from NRC (2001) [25]. The corrected positive sign for the dietary concentrate term in the dry-forage equation was used, and the equations were considered suitable empirical tools for lactating dairy cows based on the evaluation by Seo et al. [26]. Although NASEM (2021) represents the current reference for dairy cattle nutrient requirements [27], the NRC equations were retained in the present study because they provide separate passage-rate estimates for wet forages, dry forages, and concentrates. The following equations were used to calculate Kp:

For wet roughages, such as silage and fresh forage:

Kp = 3.054 + 0.614x₁

For estimating dry roughages:

Kp = 3.362 + 0.479x₁ + 0.007x₂ − 0.017x₃

For estimating concentrates:

Kp = 2.904 + 1.375x₁ − 0.020x₂

where Kp is the fractional ruminal passage rate (%/h); x₁ is DMI expressed as a percentage of BW; x₂ is the proportion of concentrate in dietary DM (%); and x₃ is the NDF concentration of the individual dry forage (% DM).

For the standardized model calculations, DMI was set at 4% of BW and the dietary concentrate proportion was set at 50% of DM. These rounded values were selected to define a simplified reference scenario for within-study comparisons among feedstuffs and were not derived from the measured intake or diet composition of the donor cows. The donor cows had a mean DMI of 26.39 kg/d and a mean BW of 514 kg, corresponding to approximately 5.13% of BW, and their basal diet contained 58.38% concentrate on a DM basis. Thus, the difference between the observed animal values and the standardized model inputs was intentional. The standardized inputs were retained to evaluate all feedstuffs under identical calculation conditions rather than to estimate animal-specific passage rates. No common dietary NDF value was imposed; for the dry-forage equation, x3 was the measured NDF concentration of each individual dry forage.

Whole-plant corn silage 1, whole-plant corn silage 2, and corn stover silage were classified as wet forages, for which the calculated Kp was 5.5%/h. Wheat straw and alfalfa were classified as dry forages. Because the NRC dry-forage equation incorporates the NDF concentration of the individual feedstuff, the calculated Kp values were 4.3%/h for wheat straw and 4.9%/h for alfalfa. The remaining concentrate-type feedstuffs and by-products were assigned a calculated Kp of 7.4%/h.

These equations were considered appropriate for the present standardized comparisons because the donor animals were multiparous, mid-lactation Holstein cows, which fall within the lactating dairy-cow population for which the equations were developed and evaluated, and because the tested feedstuffs could be classified into the corresponding wet-forage, dry-forage, or concentrate categories. However, the equations do not explicitly account for feed particle characteristics, processing effects, liquid flow, or temporal variation in intake. Moreover, the Kp values were not measured in the donor cows, and the standardized DMI and dietary concentrate inputs were not intended to reproduce their individual passage rates. Therefore, the calculated Kp values were used as standardized scenario inputs rather than as animal- or feed-specific measurements, and the resulting intestinally absorbable amino acid flow estimates should not be generalized to other animals or diets without recalculation.

Amino Acid Composition of Ruminal Microbial Protein

The amino acid composition of ruminal microbial cell-wall and non-cell-wall protein used in the CNCPS-based calculations is presented in Table A7. The means and standard deviations for ruminal microbial protein were derived from 441 microbial samples obtained from 61 diets in 35 experiments [28]. Because tryptophan content was not reported separately for the microbial cell-wall fraction, it was assumed to be identical to that of the non-cell-wall fraction [29].

Calculation Pathways for Intestinally Absorbable Essential Amino Acid Flows

All intestinally absorbable essential amino acid flows reported in this study were model-derived estimates and were not measured directly in intestinal digesta. Two calculation pathways were used. In the first pathway, measured chemical composition and insoluble-protein amino acid composition, experimentally determined CNCPS protein fractions, NRC-derived ruminal passage rates (Kp) calculated under the standardized dietary scenario, and Kd1, Kd2, and Kd3 values fitted from the in situ ruminal disappearance data were entered into the CNCPS-based calculation framework. These outputs are hereafter referred to as CNCPS-based estimates using in situ-fitted Kd values.

In the second pathway, the same measured inputs, CNCPS fractions, Kp assumptions, and ruminal nitrogen-balance scenarios were retained, but the in situ-fitted Kd1, Kd2, and Kd3 values were replaced by values estimated from the exploratory regression equations reported in Section 3.2.3. These outputs are hereafter referred to as regression-assisted CNCPS estimates using regression-estimated Kd values. The regression equations estimated the protein-fraction degradation rate constants only and did not directly predict intestinal amino acid flows.

The intestinally absorbable flow of each essential amino acid derived from rumen-undegraded feed protein was then calculated using the following equation in both calculation pathways:

REFAAij = AAINSPij × 0.01 × CPj × 10 × [(PB1j/100) × Kpj/(Kd1j + Kpj) + (PB2j/100) × Kpj/(Kd2j + Kpj) + (PB3j/100) × Kpj/(Kd3j + Kpj)]

where REFAAij is the intestinally absorbable flow of amino acid i derived from rumen-undegraded protein of feedstuff j (g/kg DM); AAINSPij is the measured content of amino acid i in the insoluble protein of feedstuff j (g/100 g insoluble protein); CPj is the measured crude protein content of feedstuff j (% DM); PB1j, PB2j, and PB3j are the experimentally determined proportions of the corresponding CNCPS protein fractions (% CP); Kpj is the NRC-derived ruminal passage rate assigned to or calculated for feedstuff j under the standardized dietary scenario (%/h); and Kd1j, Kd2j, and Kd3j are either the in situ-fitted or regression-estimated degradation rate constants used in the corresponding calculation pathway (%/h). The factor 10 converts CP from % DM to g/kg DM. Because fraction-specific amino acid compositions were not measured, the measured insoluble-protein amino acid profile of each feedstuff was applied to the rumen-undegraded portions of PB1, PB2, and PB3 as a proxy for the amino acid composition of rumen-undegraded feed protein [12].

The microbial-protein contribution was calculated within the CNCPS framework under positive ruminal nitrogen balance and ruminal nitrogen deficiency, respectively, and was combined with the rumen-undegraded feed-protein contribution to obtain total intestinally absorbable essential amino acid flow. These two nitrogen-balance conditions were predefined CNCPS calculation scenarios applied to all feedstuffs; ruminal nitrogen balance was not measured in the donor cows, and no experimental nitrogen-deficiency treatment was imposed.

2.4. Statistical Analysis

Data are presented as means ± SD unless otherwise stated. For chemical composition and CNCPS fractions, the independent feed batch was the experimental unit, and triplicate laboratory determinations were analytical replicates. For in situ disappearance measurements at each incubation time, cow was the biological replicate. However, the nonlinear degradation parameters were estimated from the mean disappearance series across the four cows for each feedstuff. Therefore, each feedstuff contributed one fitted value for each Kd parameter to the across-feedstuff regression analyses, and between-cow variability in the fitted Kd values was not quantified. Each feedstuff was evaluated in one sequential run using the same four cows; because run was not replicated, run-to-run variation could not be separated from feedstuff effects, and feedstuff comparisons were descriptive unless an inferential test and p-value were reported. Regression analyses used one feedstuff-level mean per variable (n = 19).

Before analysis and nonlinear fitting, records were checked for completeness, unit consistency, calculation accuracy, duplicate or missing entries, and biologically impossible values, including negative or >100% disappearance and CNCPS-inconsistent fractions. Suspect values were verified against laboratory records. No imputation was used; feedstuffs missing variables required for an analysis were excluded from that analysis. Feed names and units were standardized, and variables were expressed on the appropriate % DM, % CP, or % CHO basis.

GraphPad Prism version 5.0 for Windows (GraphPad Software Inc., La Jolla, CA, USA) was used for nonlinear least-squares fitting of degradation curves. SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) was used for multiple linear regression, principal component analysis, hierarchical cluster analysis, and the associated statistical diagnostics. Exploratory regressions examined (1) relationships of conventional nutrients with CNCPS carbohydrate and protein fractions and (2) relationships of conventional nutrients or CNCPS fractions with fitted Kd1, Kd2, and Kd3. Candidate predictors were DM, CP, NDF, ADF, EE, ash, ADL, starch, NPN, SCP, NDIP, ADIP, and relevant CNCPS fractions. NDIP and ADIP values in Table A3 were converted from % DM to % CP before protein-related regressions.

Predictors were screened manually for statistical association, mathematical independence, collinearity, biological plausibility, and coefficient stability across alternative models and leave-one-observation-out analyses. Biological plausibility required a reasonable relationship with protein solubility, detergent-insoluble protein, cell-wall association, readily fermentable carbohydrate availability, or ruminal accessibility. Variables were rejected when their contribution reflected only improved in-sample fit, mathematical redundancy, an unstable or biologically implausible coefficient direction, or dependence on a single feedstuff or broad feed-category contrast. Final models were selected using overall and predictor significance, in-sample fit, parsimony, biological interpretability, and coefficient stability. No automatic stepwise, forward, or backward selection was used, and no final model contained more than three predictors.

For multiple-predictor equations, VIF < 5 was required. Residual normality was evaluated using the Shapiro–Wilk test and normal Q–Q plots; p > 0.05 indicated no significant departure from normality. Residual-versus-fitted plots were used to assess linearity and homoscedasticity. Influential observations were screened using Cook’s distance > 4/n, leverage > 2(p + 1)/n, and |studentized residual| > 2, where p was the number of predictors. Flagged observations were examined individually and by leave-one-observation-out sensitivity analysis but were not automatically excluded.

Broad feed category was not included as a predictor because category sizes were small and unbalanced. Residual and influence diagnostics were nevertheless examined by category, and concentrate-type and roughage residuals were compared using Welch’s t-test. Model fit was summarized by R², RMSE, MAE, overall model p-value, and n. Statistical significance was set at p < 0.05 and high significance at p < 0.01. All regression statistics describe in-sample behavior only; because n = 19 and no independent validation dataset was available, the equations were treated as exploratory and not as practical prediction tools.

PCA and hierarchical cluster analysis were conducted using the data in Table 1, Table 2 and Table 3 for CP, NDF, ADF, EE, ash, starch, CA, CB1, CB2, CC, PA, PB1, PB2, PB3, and PC. Variables were standardized using z-scores, and clustering used Euclidean distance and Ward’s method. Because CNCPS carbohydrate and protein fractions sum to 100% within their respective pools, PCA loadings were interpreted as descriptive relative-composition patterns rather than independent fraction effects. These analyses characterized compositional heterogeneity and category-related clustering but were not used for predictor selection or regression construction. Given the small and unbalanced category sizes, separate category-specific models were not fitted.

Table 1.

Conventional chemical composition of the tested feed samples.

Feed	DM	CP	NDF	ADF	EE	Ash	ADL	Starch
Feed	DM	DM	DM	DM	DM	DM	DM	DM
Soybean meal	87.79 ± 0.55	49.45 ± 1.11	15.61 ± 1.12	10.25 ± 0.74	1.46 ± 0.13	6.59 ± 0.36	0.18 ± 0.02	8.34 ± 0.90
Wheat	87.18 ± 0.70	15.60 ± 0.62	19.16 ± 1.92	2.46 ± 0.30	2.00 ± 0.20	3.22 ± 0.19	0.46 ± 0.07	58.32 ± 2.33
DDGS	89.32 ± 0.98	32.74 ± 1.80	39.11 ± 3.01	10.38 ± 0.91	5.31 ± 0.41	4.72 ± 0.36	3.15 ± 0.35	12.20 ± 1.34
Molassed soybean hulls	91.11 ± 0.77	15.74 ± 0.66	56.13 ± 2.95	33.66 ± 2.12	4.52 ± 0.38	4.86 ± 0.31	1.86 ± 0.20	5.66 ± 0.71
Extruded soybean hulls	89.82 ± 0.68	11.33 ± 0.43	70.45 ± 3.35	45.41 ± 2.59	1.39 ± 0.11	4.93 ± 0.28	1.71 ± 0.16	3.51 ± 0.40
Wheat middlings	88.46 ± 0.97	20.73 ± 1.14	26.92 ± 2.96	5.62 ± 0.74	4.47 ± 0.39	3.27 ± 0.29	1.24 ± 0.16	40.39 ± 3.55
Wheat bran	86.73 ± 0.91	19.65 ± 1.03	47.70 ± 5.01	11.60 ± 1.46	6.98 ± 0.59	7.04 ± 0.59	3.03 ± 0.38	11.21 ± 0.94
Flaked corn	86.65 ± 0.62	9.22 ± 0.33	17.78 ± 1.60	3.37 ± 0.36	3.36 ± 0.30	1.29 ± 0.07	0.74 ± 0.10	64.29 ± 2.31
Cottonseed 1	89.68 ± 0.90	23.58 ± 0.94	45.83 ± 2.75	40.88 ± 2.86	12.50 ± 0.75	3.91 ± 0.27	15.86 ± 1.27	5.40 ± 0.65
Cottonseed 2	92.45 ± 1.02	24.78 ± 1.09	41.15 ± 2.72	32.58 ± 2.51	19.72 ± 1.30	4.58 ± 0.35	11.98 ± 1.05	5.92 ± 0.78
Pelleted sugar beet pulp	91.48 ± 0.70	12.40 ± 0.47	50.03 ± 2.38	25.28 ± 1.44	4.05 ± 0.31	4.19 ± 0.24	2.55 ± 0.24	5.57 ± 0.63
whole-plant corn silage 1	29.93 ± 1.16	9.37 ± 0.59	67.73 ± 4.98	30.23 ± 2.54	3.87 ± 0.41	5.28 ± 0.44	7.62 ± 0.80	6.02 ± 0.76
whole-plant corn silage 2	35.10 ± 1.30	8.69 ± 0.57	36.88 ± 2.84	21.56 ± 1.90	4.07 ± 0.45	4.73 ± 0.42	2.73 ± 0.30	7.38 ± 0.97
Corn stover silage	24.93 ± 1.15	9.22 ± 0.61	60.96 ± 4.69	35.74 ± 3.15	5.45 ± 0.60	8.38 ± 0.74	4.39 ± 0.48	5.51 ± 0.73
Wheat straw	90.46 ± 0.90	4.98 ± 0.33	78.65 ± 6.06	51.83 ± 4.56	1.31 ± 0.14	13.41 ± 1.18	7.89 ± 0.87	2.82 ± 0.47
Alfalfa	91.68 ± 0.83	16.32 ± 0.98	44.55 ± 3.12	34.69 ± 2.78	3.60 ± 0.36	3.52 ± 0.28	7.70 ± 0.77	5.08 ± 0.76
Dried brewer grains	93.83 ± 1.03	24.76 ± 1.36	52.79 ± 4.06	23.86 ± 2.10	12.94 ± 1.00	4.25 ± 0.33	6.58 ± 0.72	2.27 ± 0.25
Dried distiller grains 1	93.60 ± 0.98	30.33 ± 1.59	49.33 ± 3.63	21.35 ± 1.79	10.87 ± 0.80	4.39 ± 0.32	4.17 ± 0.44	5.31 ± 0.56
Dried distiller grains 2	92.80 ± 1.02	23.54 ± 1.29	53.01 ± 4.08	19.89 ± 1.75	10.01 ± 0.77	4.68 ± 0.36	4.00 ± 0.44	6.91 ± 0.76

Feed	CHO	CA	CB1	CNSC	CB2	CC
Feed	DM	CHO	CHO	CHO	CHO	CHO
Soybean meal	42.50 ± 2.42	46.84 ± 0.91	19.62 ± 0.15	66.46 ± 2.76	32.51 ± 0.95	1.03 ± 0.03
Wheat	79.18 ± 0.35	9.26 ± 0.98	73.65 ± 0.55	82.91 ± 2.56	15.70 ± 0.90	1.39 ± 0.10
DDGS	57.24 ± 0.26	19.41 ± 0.36	21.32 ± 0.35	40.73 ± 1.98	46.08 ± 0.02	13.19 ± 0.74
Molassed soybean hulls	74.88 ± 0.48	20.92 ± 0.67	7.55 ± 0.43	28.48 ± 0.24	65.57 ± 0.03	5.95 ± 0.62
Extruded soybean hulls	82.35 ± 0.45	16.44 ± 0.30	4.26 ± 0.12	20.70 ± 0.18	74.32 ± 0.83	4.98 ± 0.44
Wheat middlings	71.54 ± 0.66	8.15 ± 0.69	56.46 ± 0.47	64.61 ± 1.22	31.21 ± 0.55	4.17 ± 0.07
Wheat bran	66.33 ± 0.40	20.48 ± 0.69	16.90 ± 0.24	37.38 ± 0.44	51.65 ± 0.54	10.97 ± 1.01
Flaked corn	86.13 ± 1.13	6.97 ± 0.78	74.65 ± 0.35	81.61 ± 1.43	16.33 ± 0.35	2.06 ± 0.57
Cottonseed 1	60.01 ± 2.66	18.75 ± 0.49	8.99 ± 0.06	27.74 ± 0.56	8.85 ± 0.21	63.41 ± 0.71
Cottonseed 2	50.92 ± 0.98	16.56 ± 0.31	11.62 ± 0.18	28.18 ± 1.49	15.33 ± 0.61	56.49 ± 0.24
Pelleted sugar beet pulp	79.36 ± 1.16	36.61 ± 0.55	7.02 ± 0.33	43.64 ± 0.22	48.66 ± 0.01	7.70 ± 0.42
whole-plant corn silage 1	81.48 ± 0.36	12.03 ± 0.95	7.39 ± 0.67	19.42 ± 0.72	58.14 ± 0.14	22.44 ± 0.13
whole-plant corn silage 2	82.51 ± 0.68	47.82 ± 0.19	8.95 ± 0.15	56.77 ± 0.04	35.30 ± 0.11	7.93 ± 0.34
Corn stover silage	76.95 ± 0.53	15.21 ± 0.26	7.16 ± 0.02	22.37 ± 1.29	63.94 ± 0.23	13.69 ± 0.96
Wheat straw	80.30 ± 0.76	0.06 ± 0.02	3.51 ± 0.07	3.57 ± 0.91	72.84 ± 0.17	23.59 ± 0.74
Alfalfa	76.56 ± 0.07	36.55 ± 0.04	6.64 ± 0.48	43.19 ± 0.44	32.67 ± 0.75	24.14 ± 0.35
Dried brewer grains	58.05 ± 0.01	8.18 ± 0.72	3.91 ± 0.13	12.09 ± 0.59	60.72 ± 0.82	27.19 ± 0.14
Dried distiller grains 1	54.41 ± 1.14	7.82 ± 0.63	9.76 ± 0.01	17.58 ± 0.62	64.04 ± 0.57	18.38 ± 0.24
Dried distiller grains 2	61.77 ± 0.83	9.51 ± 0.79	11.18 ± 0.66	20.69 ± 0.87	63.77 ± 0.54	15.53 ± 0.26

Feed	PA	PB1	PB2	PB3	PC
Soybean meal	7.81 ± 0.86	6.78 ± 0.62	83.01 ± 0.48	0.46 ± 0.09	1.94 ± 0.02
Wheat	8.67 ± 0.16	24.40 ± 0.11	35.52 ± 0.01	25.52 ± 0.18	5.89 ± 0.11
DDGS	22.84 ± 0.25	3.07 ± 0.03	59.95 ± 1.83	11.31 ± 0.31	2.83 ± 0.02
Molassed soybean hulls	15.85 ± 0.28	18.19 ± 0.61	51.78 ± 0.43	8.23 ± 0.12	5.95 ± 0.03
Extruded soybean hulls	15.20 ± 0.25	9.02 ± 0.36	61.66 ± 0.73	9.44 ± 0.42	4.68 ± 0.05
Wheat middlings	14.59 ± 0.04	25.31 ± 0.51	53.41 ± 0.59	4.49 ± 0.10	2.21 ± 0.01
Wheat bran	15.13 ± 0.17	18.69 ± 0.54	38.03 ± 0.91	23.33 ± 0.03	4.81 ± 0.07
Flaked corn	9.28 ± 0.26	1.47 ± 0.01	69.73 ± 0.28	9.00 ± 0.06	10.47 ± 0.06
Cottonseed 1	23.81 ± 0.96	4.89 ± 0.68	61.75 ± 0.30	5.89 ± 0.01	3.66 ± 0.24
Cottonseed 2	12.44 ± 0.13	5.62 ± 0.05	64.43 ± 0.08	14.41 ± 0.51	3.11 ± 0.03
Pelleted sugar beet pulp	12.69 ± 0.03	0.08 ± 0.01	48.20 ± 0.12	35.62 ± 0.50	3.47 ± 0.50
whole-plant corn silage 1	69.55 ± 0.03	2.46 ± 0.04	7.98 ± 0.01	9.28 ± 0.07	10.73 ± 0.07
whole-plant corn silage 2	63.35 ± 0.16	4.17 ± 0.11	19.69 ± 0.04	1.37 ± 0.14	11.41 ± 0.14
Corn stover silage	61.29 ± 0.02	2.02 ± 0.03	24.20 ± 0.01	1.59 ± 0.15	10.90 ± 0.11
Wheat straw	55.06 ± 0.26	8.03 ± 0.08	14.02 ± 0.03	4.35 ± 0.01	18.54 ± 0.02
Alfalfa	17.59 ± 0.55	12.40 ± 0.54	59.29 ± 0.21	4.84 ± 0.01	5.88 ± 0.04
Dried brewer grains	73.45 ± 0.28	7.72 ± 0.12	12.35 ± 0.56	2.92 ± 0.02	3.57 ± 0.02
Dried distiller grains 1	4.79 ± 0.06	2.49 ± 0.02	78.88 ± 0.69	10.58 ± 0.02	3.27 ± 0.06
Dried distiller grains 2	3.14 ± 0.13	0.76 ± 0.01	80.24 ± 0.21	11.67 ± 0.07	4.19 ± 0.06

Feed	Kd1 (%/h)	Kd2 (%/h)	Kd3 (%/h)	R²
Soybean meal	252.90	18.73	1.78	0.839
Wheat	93.79	9.41	6.74	0.762
DDGS	159.50	6.15	2.03	0.798
Molassed soybean hulls	236.70	5.09	2.54	0.970
Extruded soybean hulls	145.10	2.29	1.85	0.777
Wheat middlings	250.20	8.05	7.19	0.893
Wheat bran	396.20	14.91	5.92	0.805
Flaked corn	122.40	8.40	8.16	0.822
Cottonseed 1	37.05	2.66	2.25	0.760
Cottonseed 2	46.93	8.47	6.76	0.887
Pelleted sugar beet pulp	177.50	11.96	2.45	0.890
whole-plant corn silage 1	137.50	9.41	6.92	0.862
whole-plant corn silage 2	177.00	14.31	6.60	0.978
Corn stover silage	88.71	11.79	1.62	0.646
Wheat straw	86.42	5.21	4.47	0.815
Alfalfa	214.50	12.28	1.35	0.796
Dried brewer grains	101.80	4.73	2.05	0.884
Dried distiller grains 1	145.60	2.98	2.07	0.687
Dried distiller grains 2	113.40	9.42	4.48	0.870

Feed	Kd4 (%/h)	R²	Kd5 (%/h)	R²	Kd6 (%/h)	R²
Soybean meal	56.49	0.999	47.18	0.977	1.739	0.767
Wheat	128.9	0.996	13.84	0.941	0.2865	0.030 ⁺
DDGS	118.6	0.996	2.757	0.979	2.291	0.956
Molassed soybean hulls	60.30	0.967	5.108	0.950	1.836	0.955
Extruded soybean hulls	32.76	0.992	5.331	0.975	1.293	0.852
Wheat middlings	28.63	0.988	16.45	0.926	2.413	0.921
Wheat bran	101.1	0.540	8.395	0.920	1.875	0.863
Flaked corn	214.6	0.722	5.731	0.994	2.034	0.914
Cottonseed 1	16.89	0.949	1.004	0.745	2.955	0.869
Cottonseed 2	20.13	0.831	2.592	0.949	1.929	0.269 ⁺
Pelleted sugar beet pulp	26.67	0.716	3.854	0.953	4.513	0.930
whole-plant corn silage 1	28.40	0.442 ⁺	26.48	0.841	1.175	0.647
whole-plant corn silage 2	32.00	0.498 ⁺	18.55	0.876	1.872	0.680
Corn stover silage	150.8	0.936	40.06	0.936	2.172	0.994
Wheat straw	50.04	0.941	2.676	0.907	1.51	0.955
Alfalfa	32.49	0.761	10.29	0.838	1.529	0.970
Dried brewer grains	28.82	0.724	9.032	0.834	1.874	0.778
Dried distiller grains 1	34.52	0.714	11.69	0.878	1.773	0.789
Dried distiller grains 2	3.160	0.996	2.124	0.984	1.762	0.952

Feed	Lys	Met	Total EAA
Soybean meal	2.911	0.261	24.33
Wheat	0.792	0.333	7.065
DDGS	0.719	0.366	14.71
Molassed soybean hulls	0.649	0.218	5.795
Extruded soybean hulls	0.533	0.249	5.500
Wheat middlings	0.486	0.242	6.645
Wheat bran	0.505	0.267	6.983
Flaked corn	0.166	0.247	4.906
Cottonseed 1	0.702	0.344	10.02
Cottonseed 2	0.701	0.256	10.71
Pelleted sugar beet pulp	0.700	0.277	6.047
whole-plant corn silage 1	0.069	0.24	3.087
whole-plant corn silage 2	0.078	0.233	3.159
Corn stover silage	0.109	0.243	3.388
Wheat straw	0.051	0.297	3.494
Alfalfa	0.774	0.25	7.786
Dried brewer grains	0.239	0.279	5.794
Dried distiller grains 1	0.782	0.32	10.186
Dried distiller grains 2	0.648	0.264	11.077

Item	Soybean meal	Wheat	DDGS	Molassed Soybean hulls	Extruded Soybean hulls	Wheat middlings	Alfalfa
Met	3.598	1.003	1.149	0.770	0.351	1.055	1.039
Lys	13.52	2.969	2.990	2.346	0.966	3.080	3.180
His	4.901	1.023	1.611	0.790	0.320	1.119	1.051
Phe	11.19	2.365	6.180	1.971	1.063	2.561	2.300
Trp	3.116	0.578	0.708	0.524	0.506	0.678	0.648
Thr	9.017	1.986	2.570	1.539	0.583	2.128	2.175
Leu	13.64	2.696	5.807	2.126	0.821	2.999	2.983
Ile	9.846	2.075	2.715	1.616	0.621	2.241	2.293
Val	12.25	2.767	5.605	2.350	1.218	3.102	2.811
Arg	12.18	2.407	3.030	1.910	0.719	2.802	2.674
TEAAA	93.26	19.87	32.37	15.94	7.168	21.77	21.15
Item	Wheat bran	Flaked corn	Cottonseed 1	Cottonseed 2	Pelleted sugar beet pulp	Whole-plant corn silage 1
Met	1.190	0.482	0.878	0.674	0.619	0.133
Lys	3.497	1.279	2.275	1.988	1.821	0.366
His	1.264	0.463	1.046	0.849	0.623	0.122
Phe	2.695	1.215	3.230	3.541	1.552	0.292
Trp	0.722	0.283	0.582	0.496	0.348	0.079
Thr	2.382	0.902	1.586	1.394	1.212	0.251
Leu	3.351	1.366	2.327	2.079	1.619	0.343
Ile	2.536	0.940	1.629	1.455	1.244	0.265
Val	3.283	1.425	4.284	3.327	1.965	0.341
Arg	3.172	1.108	3.214	2.595	1.423	0.310
TEAA	24.09	9.466	21.05	18.40	12.43	2.500
Item	Whole-plant corn silage 2	Corn stover silage	Wheat straw	Dried brewer grains	Dried distiller grains 1	Dried distiller grains 2
Met	0.178	0.194	0.099	0.404	1.056	0.639
Lys	0.512	0.541	0.267	1.104	2.888	1.825
His	0.169	0.179	0.087	0.429	1.345	0.754
Phe	0.376	0.428	0.225	0.927	3.994	3.278
Trp	0.107	0.114	0.057	0.233	0.774	0.650
Thr	0.350	0.370	0.182	0.791	2.227	1.363
Leu	0.477	0.504	0.247	1.112	4.033	2.297
Ile	0.369	0.391	0.192	0.816	2.571	1.548
Val	0.442	0.502	0.266	1.196	5.482	3.261
Arg	0.434	0.457	0.225	0.972	3.135	1.739
TEAA	3.416	3.679	1.847	7.985	27.51	17.36

Item	Soybean meal	Wheat	DDGS	Molassed Soybean hulls	Extruded Soybean hulls	Wheat middlings	Alfalfa
Met	1.182	0.189	0.673	0.161	0.105	0.172	0.113
Lys	7.733	0.466	1.472	0.480	0.244	0.377	0.349
His	3.303	0.208	1.187	0.180	0.082	0.233	0.123
Phe	9.272	0.854	5.734	0.847	0.475	0.865	0.558
Trp	2.241	0.079	0.420	0.159	0.270	0.142	0.087
Thr	4.945	0.275	1.598	0.261	0.113	0.286	0.246
Leu	9.150	0.401	4.781	0.415	0.179	0.525	0.399
Ile	5.796	0.274	1.694	0.272	0.124	0.304	0.264
Val	9.249	0.955	4.862	1.007	0.532	1.079	0.739
Arg	7.728	0.269	1.799	0.318	0.136	0.510	0.268
TEAA	60.60	3.972	24.22	4.100	2.262	4.493	3.146
Item	Wheat bran	Flaked corn	Cottonseed 1	Cottonseed 2	Pelleted sugar beet pulp	Whole-plant corn silage 1
Met	0.194	0.154	0.349	0.309	0.230	0.059
Lys	0.399	0.158	0.759	0.859	0.594	0.139
His	0.290	0.121	0.474	0.509	0.233	0.047
Phe	0.925	0.756	1.829	3.176	0.990	0.150
Trp	0.117	0.078	0.248	0.287	0.096	0.033
Thr	0.270	0.156	0.543	0.630	0.360	0.096
Leu	0.568	0.456	0.878	1.084	0.470	0.135
Ile	0.325	0.150	0.542	0.650	0.332	0.101
Val	1.193	0.862	2.506	2.755	1.353	0.171
Arg	0.616	0.170	1.609	1.773	0.317	0.117
TEAA	4.896	3.060	9.737	12.03	4.977	1.049
Item	Whole-plant corn silage 2	Corn stover silage	Wheat straw	Dried brewer grains	Dried distiller grains 1	Dried distiller grains 2
Met	0.055	0.061	0.041	0.174	0.844	0.354
Lys	0.134	0.136	0.083	0.370	2.110	0.883
His	0.045	0.046	0.026	0.203	1.239	0.527
Phe	0.140	0.173	0.119	0.517	4.321	3.812
Trp	0.032	0.033	0.021	0.090	0.694	0.615
Thr	0.093	0.094	0.057	0.300	1.792	0.782
Leu	0.131	0.133	0.078	0.463	3.840	1.765
Ile	0.098	0.100	0.059	0.295	2.201	0.999
Val	0.160	0.197	0.139	0.726	6.139	3.553
Arg	0.113	0.114	0.068	0.357	2.730	1.039
TEAA	1.001	1.087	0.693	3.494	25.909	14.33

DM	Dry matter
CP	Crude protein
NDF	Neutral detergent fiber
ADF	Acid detergent fiber
EE	Ether extract
ADL	Acid detergent lignin
NDIP	Neutral detergent insoluble protein
ADIP	Acid detergent insoluble protein
NPN	Non-protein nitrogen
SCP	Soluble crude protein
CNSC	Non-Structural Carbohydrates
CHO	Total carbohydrate
CA	Rapidly fermentable non-starch carbohydrate fraction
CB1	Starch fraction
CB2	Potentially available cell-wall carbohydrate fraction
CC	Unavailable cell-wall carbohydrate fraction
PA	Non-protein nitrogen fraction
PB1	Rapidly degradable true protein fraction
PB2	Moderately degradable true protein fraction
PB3	Slowly degradable true protein fraction
PC	Unavailable protein fraction
EAA	Essential amino acids
NEAA	Non-essential amino acids
NEL	Net energy for lactation
TAA	Total amino acids
Met	Methionine
Lys	Lysine
His	Histidine
Phe	Phenylalanine
Trp	Tryptophan
Thr	Threonine
Leu	Leucine
Ile	Isoleucine
Val	Valine
Arg	Arginine
TEAA	Total essential amino acids

Feed	Source/ Location	Collection or Purchase Period	Feed Type and Key Agronomic/Processing Characteristics	Storage
Soybean meal	Shandong, China	October–November 2024	Solvent-extracted protein concentrate produced after soybean oil extraction and dried in meal form.	D
Wheat	Shandong, China	June 2024	Mature cereal grain harvested at full maturity, cleaned, dried, and ground before analysis.	D
Flaked corn	Shandong, China	October–November 2024	Mature corn grain steam-conditioned at approximately 95–100 °C for approximately 30 min and roller-flaked to a density of approximately 0.32–0.39 kg/L.	D
Wheat bran	Shandong, China	October–November 2024	Wheat milling by-product consisting primarily of the outer bran fraction.	D
Molassed soybean hulls	Beijing, China	October–November 2024	Soybean hulls mixed with molasses to improve palatability and fermentable carbohydrate supply.	D
Extruded soybean hulls	Beijing, China	October–November 2024	Soybean hulls subjected to extrusion processing and used as a digestible-fiber by-product feed.	D
Pelleted sugar beet pulp	Tianjin, China	October–November 2024	Sugar beet pulp dried and pelleted after sugar extraction.	D
DDGS	Tianjin, China	October–November 2024	Distillers dried grains with solubles produced after grain fermentation, distillation, and drying.	D
Wheat middlings	Tianjin, China	October–November 2024	Wheat milling by-product consisting mainly of fine bran, germ, and residual endosperm particles.	D
Cottonseed 1	Shanxi, China	September–November 2024	Independent whole-cottonseed commercial samples obtained from Shanxi.	D
Cottonseed 2	Tianjin, China	September–November 2024	Independent whole-cottonseed commercial samples obtained from Tianjin.	D
Whole-plant corn silage 1	Beijing Sunlon Livestock Development Co., Ltd., Beijing, China	Harvested September–October 2024; sampled November–December 2024	hole corn plants harvested at the late milk to early dent stage, chopped to approximately 1–2 cm, compacted, and anaerobically ensiled.	S
Whole-plant corn silage 2	Tianjin Jialihe Animal Husbandry Group Co., Ltd., Tianjin, China	Harvested September–October 2024; sampled November–December 2024	Whole corn plants harvested at the half-milk-line to dent stage, chopped to approximately 1–2 cm, compacted, and anaerobically ensiled.	S
Corn stover silage	Tianjin, China	Prepared September–October 2024; sampled November–December 2024	Corn stover collected after grain harvest, chopped, compacted, and anaerobically ensiled	S
Wheat straw	Henan, China	June 2024	Wheat straw collected after grain harvest, air-dried, chopped, and ground before analysis.	D
Alfalfa	Shandong, China	Late May 2024	First-cut alfalfa harvested at the early flowering stage in late May 2024, sun-cured as hay, and subsequently sampled.	D
Dried brewer grains	Henan, China	October–November 2024	Brewer grains generated after mashing and lautering during beer production and subsequently dried.	D
Dried distiller grains 1	Shanxi Chongkang, Shanxi, China	October–November 2024	Independent commercial dried distiller grains obtained from Shanxi Chongkang after grain fermentation and distillation.	D
Dried distiller grains 2	Shanxi Ruimei, Shanxi, China	October–November 2024	Independent commercial dried distiller grains obtained from Shanxi Ruimei after grain fermentation and distillation.	D

Ingredient	Content, % DM	Nutrient Level	Content
Whole-plant corn silage	28.15	NEL, MJ/kg DM	7.11
Alfalfa	13.47	CP, % DM	17.96
Corn meal	11.47	EE, % DM	4.54
Flaked corn	11.77	NDF, % DM	35.88
Soybean meal	15.31	ADF, % DM	21.90
Corn gluten meal	1.68	Ca, % DM	1.03
Double-low rapeseed meal	2.88	P, % DM	0.42
Whole cottonseed	2.93
Distillers’ grains	6.80
Palm oil	1.21
NaHCO₃	1.35
Limestone	0.55
CaHPO₄	0.52
MgCl₂	0.39
NaCl	0.52
Premix	1.00
Total	100.00

Feed	NDIP	ADIP	NPN	SCP
Feed	DM	DM	SCP	CP
Soybean meal	1.35 ± 0.12	1.09 ± 0.10	53.51 ± 3.85	14.59 ± 0.92
Wheat	5.62 ± 0.56	1.05 ± 0.11	26.23 ± 2.62	33.07 ± 2.65
DDGS	5.18 ± 0.46	1.04 ± 0.09	88.14 ± 6.79	25.91 ± 2.85
Molassed soybean hulls	2.57 ± 0.22	1.08 ± 0.11	46.56 ± 3.91	34.04 ± 2.86
Extruded soybean hulls	1.60 ± 0.12	0.53 ± 0.05	62.75 ± 4.77	24.22 ± 1.84
Wheat middlings	6.17 ± 0.68	1.05 ± 0.12	36.57 ± 4.02	39.89 ± 3.51
Wheat bran	1.95 ± 0.20	1.04 ± 0.11	44.74 ± 4.70	33.83 ± 2.84
Flaked corn	2.47 ± 0.22	0.95 ± 0.09	86.34 ± 7.77	10.74 ± 0.77
Cottonseed 1	4.58 ± 0.37	0.81 ± 0.06	82.95 ± 5.81	28.70 ± 2.58
Cottonseed 2	5.30 ± 0.47	0.47 ± 0.04	68.87 ± 5.30	18.06 ± 1.79
Pelleted sugar beet pulp	2.07 ± 0.16	1.11 ± 0.11	99.78 ± 7.58	12.72 ± 0.97
whole-plant corn silage 1	1.21 ± 0.10	1.08 ± 0.09	96.58 ± 2.03	72.01 ± 3.78
whole-plant corn silage 2	1.23 ± 0.11	1.07 ± 0.09	93.82 ± 2.06	67.53 ± 3.71
Corn stover silage	1.22 ± 0.11	0.99 ± 0.09	96.80 ± 2.13	63.32 ± 3.48
Wheat straw	1.06 ± 0.12	1.05 ± 0.12	87.27 ± 7.68	63.09 ± 6.94
Alfalfa	1.75 ± 0.18	0.96 ± 0.10	58.66 ± 4.69	29.99 ± 3.00
Dried brewer grains	4.49 ± 0.40	1.06 ± 0.09	90.49 ± 6.97	81.16 ± 8.93
Dried distiller grains 1	4.02 ± 0.34	1.06 ± 0.09	65.81 ± 4.84	7.27 ± 0.76
Dried distiller grains 2	5.20 ± 0.46	1.09 ± 0.10	80.60 ± 6.21	3.90 ± 0.43

Model	Predictor	VIF	Maximum VIF	Shapiro–Wilk W	p-Value	Feed-Category Comparison p-Value
Kd1	NDIP	1.173	1.260	0.954	0.456	0.150
	PB1	1.260
	CC	1.087
Kd2	ADF	1.370	1.385	0.957	0.515	0.159
	Ash	1.385
	CA	1.015
Kd3	Starch	1.000	1.000	0.862	0.011	0.419

Model	Feedstuff	Cook’s Distance	Leverage	Studentized Residual	R² After Removal	Overall p-Value After Removal	Main Finding
Kd1	Wheat	0.826	0.482	−2.085	0.792	<0.001	Coefficient directions unchanged
Kd1	Wheat bran	0.709	0.296	3.376	0.732	<0.001	PB1 coefficient decreased but remained significant
Kd1	Cottonseed 1	0.001	0.461	0.058	0.702	<0.001	CC became marginally nonsignificant, p ≈ 0.060
Kd2	Wheat straw	0.447	0.713	-0.841	0.740	<0.001	Coefficient directions and significance unchanged
Kd2	Soybean meal	0.003	0.450	−0.122	0.630	0.003	Model fit decreased; conclusions unchanged
Kd2	DDGS	0.140	0.113	−2.421	0.811	<0.001	Model fit increased; conclusions unchanged
Kd3	Flaked corn	0.001	0.459	0.043	0.261	0.030	Starch coefficient remained positive; R² decreased
Kd3	Wheat	0.097	0.369	−0.563	0.352	0.010	Positive starch association retained
Kd3	Whole-plant corn silage 1	0.115	0.062	2.024	0.468	0.002	Positive starch association retained

Parameter	Analytical Principle or Procedure	Standard or Methodological Basis
Dry matter	Oven drying to constant weight	GB/T 6435-2014
Crude protein	Kjeldahl determination of total nitrogen; nitrogen concentration multiplied by 6.25	GB/T 6432-2018
Ether extract	Solvent extraction of lipid-soluble components	GB/T 6433-2006
Ash	Combustion in a muffle furnace to remove organic matter	GB/T 6438-2007
Neutral detergent fiber	Neutral-detergent extraction of the insoluble cell-wall residue	GB/T 20806-2022
Acid detergent fiber	Acid-detergent extraction of cellulose, lignin, and acid-insoluble residue	NY/T 1459-2022
Lignin	Sulfuric acid treatment of the acid-detergent fiber residue	Section Determination of Conventional Nutrients; Van Soest et al. [20]
Starch	Perchloric acid extraction followed by anthrone colorimetric determination	Section Determination and Calculation of CNCPS Fractions
Non-protein nitrogen	Selective precipitation of true protein, followed by determination of nitrogen remaining in the soluble fraction	Section Determination and Calculation of CNCPS Fractions; Van Soest et al. [20]
Soluble crude protein	Buffer extraction of the soluble protein fraction, followed by nitrogen determination	Section Determination and Calculation of CNCPS Fractions; Van Soest et al. [20]
Neutral detergent insoluble protein	Determination of nitrogen remaining in the neutral detergent fiber residue	Section Determination and Calculation of CNCPS Fractions; Van Soest et al. [20]
Acid detergent insoluble protein	Determination of nitrogen remaining in the acid detergent fiber residue	Section Determination and Calculation of CNCPS Fractions; Van Soest et al. [20]

Amino Acid	Cell Wall	Non-Cell Wall	Ruminal Microbes
Amino Acid	Cell Wall	Non-Cell Wall	Mean	SD
Met	2.40	2.68	2.60	0.70
Lys	5.60	8.20	7.90	0.90
His	1.74	2.69	2.00	0.40
Phe	4.20	5.16	5.10	0.30
Trp	1.63	1.63	-	-
Thr	3.30	5.59	5.80	0.50
Leu	5.90	7.51	8.10	0.80
Ile	4.00	5.88	5.70	0.40
Val	4.70	6.16	6.20	0.60
Arg	3.82	6.96	5.10	0.70

Response Variable	Regression Equation	R²	p-Value
PB2 (% CP)	PB2 = 19.423 + 1.449 CP (% DM)	0.408	<0.01
PC (% CP)	PC = 7.423 − 0.272 CP (% DM) + 0.811 Ash (% DM)	0.781	<0.01
CA (% CHO)	CA = 5.698 + 0.632 CHO (% DM)− 1.266 ADF (% DM)	0.675	<0.01
CB1 (% CHO)	CB1 = 13.436 + 0.743 CHO (% DM) − 0.693 NDF (% DM) − 0.588 ADF (% DM)	0.763	<0.01
CB2 (% CHO)	CB2 = −5.072 + 1.615 NDF (% DM) − 0.993 ADF (% DM)	0.754	<0.01
CC (% CHO)	CC = −10.891 + 0.505 ADF (% DM) + 2.532 EE (% DM)	0.762	<0.01

PERMALINK

CNCPS-Based Nutrient Fractionation, Ruminal Degradation Kinetics, and Intestinal Amino Acid Flow Estimates in Dairy Feedstuffs

Yansong Ge

Ziyao Wang

Xu Sun

Xueyan Lin

Zhonghua Wang

Roles

Simple Summary

Abstract

1. Introduction

2. Materials and Methods

2.1. Evaluation of Carbohydrate and Protein Fractions in Common Ruminant Feedstuffs

2.1.1. Experimental Materials and Feed Sampling

2.1.2. Determination Parameters

2.1.3. Major Reagents

2.1.4. Main Equipment

2.1.5. Analytical Methods for Nutrients and CNCPS Fractions

Determination of Conventional Nutrients

Determination and Calculation of CNCPS Fractions

2.2. Evaluation of Ruminal Degradation Rate Constants of Feed Carbohydrate and Protein Fractions

2.2.1. Experimental Materials

2.2.2. Experimental Animals and Management

2.2.3. In Situ Nylon Bag Procedure

2.2.4. Experimental Instruments and Reagents

2.2.5. Analytical Methods

2.2.6. Calculation Methods

2.3. Estimation of Intestinally Absorbable Amino Acid Flows Based on the CNCPS Model

2.3.1. Feed Samples for Amino Acid Analysis

2.3.2. Amino Acid Analysis Instruments

2.3.3. Calculation of Intestinally Absorbable Amino Acid Flows

Calculation Method of Rumen Passage Rate

Amino Acid Composition of Ruminal Microbial Protein

Calculation Pathways for Intestinally Absorbable Essential Amino Acid Flows

2.4. Statistical Analysis

Table 1.

Table 2.

Table 3.

3. Results

3.1. Evaluation of Carbohydrate and Protein Fractions of Common Feeds for Ruminants

3.1.1. Proximate Composition Analysis of Common Feeds for Dairy Cows

3.1.2. Carbohydrate Content and Proportions of Carbohydrate Fractions in Feeds

3.1.3. Proportions of Crude Protein Fractions in Feeds

3.1.4. Exploratory Regression Relationships Between Conventional Nutrients and CNCPS Fractions

3.1.5. Exploratory Multivariate Analysis of Feed Nutritional Characteristics

3.2. Evaluation of Rumen Degradation Rates of Carbohydrate and Protein Fractions in Feeds

3.2.1. Study on Rumen Degradation Rate Constants of Protein Fractions in Common Dairy Feeds

Table 4.

3.2.2. Study on Rumen Degradation Rate Constants of Carbohydrate Fractions in Common Dairy Cow Feeds

Table 5.

3.2.3. Exploratory Regression Relationships Between Feed Characteristics and Protein-Fraction Degradation Rate Constants

3.3. CNCPS-Based Estimation of Intestinally Absorbable Amino Acid Flows

3.3.1. Amino Acid Composition of Insoluble Crude Protein in Feeds

Table 6.

3.3.2. Comparison of Intestinally Absorbable Essential Amino Acid Flow Estimates Using In Situ-Fitted and Regression-Estimated Kd Values

Table 7.

Table 8.

Table 9.

Table 10.

4. Discussion

4.1. Overall Contribution and Scope of the Study

4.2. Evaluation of Carbohydrate and Protein Fractions in Common Feeds for Ruminants

4.2.1. Conventional Nutrient Composition

4.2.2. Carbohydrate Fractions

4.2.3. Protein Fractions

4.2.4. Relationships Between Conventional Nutrients and CNCPS Fractions

4.3. Ruminal Degradation Kinetics of Carbohydrate and Protein Fractions

4.3.1. Protein-Fraction Degradation Constants

4.3.2. Carbohydrate-Fraction Degradation Constants

4.3.3. Regression Relationships Between Conventional Nutrients, Feed Fractions, and Ruminal Degradation Rate Constants of Protein Fractions

4.4. Interpretation of CNCPS-Based Intestinally Absorbable Amino Acid Flow Estimates

4.4.1. Insoluble-Protein Amino Acid Composition

4.4.2. Amino Acid Flow Estimates from Alternative Kd Inputs

5. Conclusions

Abbreviations

Supplementary Materials

Appendix A

Table A1.

Table A2.

Table A3.