Quantitation of relationship and development of nutrient prediction with vibrational molecular structure spectral profiles of feedstocks and co-products from canola bio-oil processing

Objective This program aimed to reveal the association of feed intrinsic molecular structure with nutrient supply to animals from canola feedstocks and co-products from bio-oil processing. The special objective of this study was to quantify the relationship between molecular spectral feature and nutrient availability and develop nutrient prediction equation with vibrational molecular structure spectral profiles. Methods The samples of feedstock (canola oil seeds) and co-products (meals and pellets) from different bio-oil processing plants in Canada (CA) and China (CH) were submitted to this molecular spectroscopic technique and their protein and carbohydrate related molecular spectral features were associated with the nutritional results obtained through the conventional methods of analyses for chemical and nutrient profiles, rumen degradable and intestinal digestible parameters. Results The results showed that the spectral structural carbohydrates spectral peak area (ca. 1,487.8 to 1,190.8 cm−1) was the carbohydrate structure that was most significant when related to various carbohydrate parameters of canola meals (p<0.05, r>0.50). And spectral total carbohydrate area (ca. 1,198.5 to 934.3 cm−1) was most significant when studying the various carbohydrate parameters of canola seeds (p<0.05, r>0.50). The spectral amide structures (ca. 1,721.2 to 1,480.1 cm−1) were related to a few chemical and nutrient profiles, Cornell Net Carbohydrate and Protein System (CNCPS) fractions, truly absorbable nutrient supply based on the Dutch protein system (DVE/OEB), and NRC systems, and intestinal in vitro protein-related parameters in co-products (canola meals). Besides the spectral amide structures, α-helix height (ca. 1,650.8 to 1,643.1 cm−1) and β-sheet height (ca. 1,633.4 to 1,625.7 cm−1), and the ratio between them have shown to be related to many protein-related parameters in feedstock (canola oil seeds). Multi-regression analysis resulted in moderate to high R2 values for some protein related equations for feedstock (canola seeds). Protein related equations for canola meals and carbohydrate related equations for canola meals and seeds resulted in weak R2 and low p values (p<0.05). Conclusion In conclusion, the attenuated total reflectance Fourier transform infrared spectroscopy vibrational molecular spectroscopy can be a useful resource to predict carbohydrate and protein-relates nutritional aspects of canola seeds and meals.


INTRODUCTION
The dairy production system, especially in Canada, uses canola meal, rather than the seeds a source of protein because canola seeds are largely crushed for its oil content generating the meal as a co-product. The literature indicates that changes in temperature and time of harvesting can alter the chemical composition of canola seeds and different processing methods can alter the composition of canola meals [1]. Furthermore, the chemical composition of feedstuffs is indispensable in animal nutrition for feeds account for around 60-75% of the costs in animal production.
Wet laboratory analyses methods and in vivo studies require intensive labor, high amounts of samples and long hours. And each day the industry brings forth a new variety of plant, a different method of processing etc. and all can affect the final product that is consumed by the animals, therefore determining the chemical composition is required and the faster this information can be obtained, the faster the industry can improve, and better animal performance and increased profits can be observed. Therefore, a fast method of analysis for canola seeds and meals would be helpful in the dairy industry, saving time in analysis and money in manipulating diets that are taking into consideration the real specific characteristics of the ingredients being used.
As an alternative to time consuming wet laboratory analysis, different infrared spectroscopy methods have gained space in animal nutrition [2,3]. Spectroscopy is being used because it studies matters though its interaction with light quickly and without damaging the sample. The attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR) analyzes the interaction of matter with infrared light on the mid-infrared region (ca. 4,000 to 800 cm -1 ) in a quick and non-destructive manner [2]. This is different from the wet analyses that use chemicals and procedures that can damage structures and alter the composition and digestibility of feeds [4][5][6].
ATR-FTIR can help us learn not only about the composition of an ingredient but also what kind of response that ingredient has when fed to a ruminant. Therefore, to understand how the intrinsic molecular structures of canola seeds and meals relate to the chemical composition, energy profile, degradability, and digestibility in the gastrointestinal tract of dairy cows is an advantage for the industry and was the aim of this study.
This large program aimed to reveal the association of feed intrinsic molecular structure with nutrient supply to animals from canola feedstocks and co-products from bio-oil processing. The special objective of this current study was to quantify the relationship between molecular spectral feature and nutrient availability and develop nutrient prediction equation with vibrational molecular structure spectral profiles.

MATERIALS AND METHODS
The University of Saskatchewan Animal Care Committee approved the animal trial under the Animal Use Protocol No. 19910012 and animals were cared for and handled in accordance with the Canadian Council of Animal Care (CCAC, 1993) regulations.

Sampling
Samples from canola bio-oil processing plants in Canada and China were collected by the Canola Council of Canada. Five different companies in Canada provided samples from seeds used and meals produces in five different batches. As well as five different companies in China provided samples from the seeds and meals from five different batches. Each company's quality control laboratory provided the samples that were later analyzed at the University of Saskatchewan in Canada.

Chemical analysis, degradation, and intestinal digestion and nutrient supply
Chemical analyses were followed the AOAC official methods of analysis [7]; energy values were determined using a chemical summary approach in NRC [8,9]; protein and carbohydrate fractions were carried out using CNCPS 6.5 system [10,11]; the in situ study required the rumen incubation of 7 g samples at 0, 2, 4, 8, 12, 24, and 48 h in four Holstein cows followed the animal care guidelines and approved by the ethics committee of the University of Saskatchewan; the in vitro study followed the three-step procedure by Calsamiglia and Stern [12]. The true nutrient supply was studied using NRC [8,9] and the Dutch protein system (DVE/OEB) [13,14]. The spectral analysis was carried out by using the ATR-FTIR technique to study carbohydrate and protein-related molecular structures. All procedures and analyses were reported in detail in the previous studies.

Statistical analysis
To study the relationship between the various spectral features to the chemical and energy profiles, and rumen and intestinal availability and digestibility, the data of interest were analyzed using the procedure CORR on SAS 9.4 (SAS Institute, USA).
The procedure REG on SAS 9.4 (SAS Institute, USA) was used for the multi-regression analysis to create prediction equations based on the data collected during this study. Only the significant model equations are represented here (R 2 > 0.60).

RESULTS AND DISCUSSION
Relationship study on carbohydrate-related spectral features and chemical and nutrient profiles and rumen degradation and intestinal digestion In a correlation analysis, there is a linear relationship between the variables analyzed (p<0.05), and the r value will determine the strength of this relationship, where r = 1 or -1, is a perfect relationship; r = 0.8 or -0.8, indicate a strong relationship; r = 0.6 or -0.6, indicate a moderate relationship; r = 0, indicates absence of linear relationship [15]. Represented from Table 1 to 9 are the correlation between carbohydraterelated molecular structures and variables from chemical and energy profiles, Cornell Net Carbohydrate and Protein System (CNCPS) fractions, NRC, and DVE/OEB systems that showed significance (p<0.05).
The structural carbohydrates spectral peak area (STCA) is the molecular structure that seems to have linear relationships with many of the characteristics of canola meals studied. It is related to the contents of cellulose and lignin (Table 1), digestible fiber fractions (Table 3), effective degradability of protein and microbial protein (MP) synthesized in the rumen based on energy (Table 5), total digestible neutral detergent fiber (tdNDF) and feed milk value (Table 6), and with endogenous crude protein (ECP) and ECP truly absorbed in the small intestine (AECP) ( Table 8). While other structures were also related, STCA was related to at least one aspect of each studied profile or system.
The total carbohydrate area (TCA) is the carbohydraterelated structure that was found to be related to many characteristics of canola seeds in this study. It is related to the sugar content (Table 2), rumen degradable and undegradable fractions of water-soluble carbohydrates (RDCA4 and RUCA4) ( Table 4), and to ECP and AECP ( Table 9). The cellulosic compounds area was linearly related to neutral detergent aspects (NDF and tdNDF) ( Tables 2 and 7). None of the carbohydrate molecular structures studied on this project appeared to have a linear relationship with any of the DVE/OEB system variables for canola seeds (p>0.05,  [2]) rereported the relationship between spectral feature collected from different spectroscopic techniques DRIFT, Synchrotron IMS, ATR-FTIR molecular spectroscopy and nutrition. However, there was no systematic relationship study in canola seeds and meal in a large scale. This current study shows that it is possible to relate many feed characteristics and ruminal and intestinal responses of canola seeds and meals fed to dairy cows from their carbohydrate-related spectral profiles revealed through molecular analysis using the ATR/FTIR technology.

Relationship study on protein-related spectral features and chemical and nutrient profiles and rumen degradation and intestinal digestion of canola seeds and meals
The linear relationship study between protein structures revealed through the ATR/FTIR technique and chemical characteristics of canola meals and seeds that were significant are presented from Table 10 to 19. Many variables from canola meals and seeds showed to be related with amides areas and heights in the present study. Weak correlations are not represented on these tables.
Amides peak area and amides height are related to the soluble crude protein content of both canola meals and seeds (Tables 10, 11), but stronger relationships were observed on canola seeds (r = 0.64, p<0.001, for peak area; r = 0.62, p<0.001, for amide height) (Table 11). Slowly degradable protein frac-   tions (PB2 and RDPB2) of canola meals are negatively related to Amide area, and moderately degradable protein (PB1) is also negatively related to amide area ratio and amide height in canola meals (Table 12). Only the truly digested protein in the small intestine (DVE) (r = 0.57, p = 0.026) and estimated milk production (DVE FMV) (r = 0.56, p = 0.028) of canola meals seemed to be related to the height of amide II (Table 14). Strong relationships can be observed between many protein structures and soluble protein fractions (PA2, RDPA2, and RUPA2), moderately degradable fractions (PB1, RDPB1, and RUPB1), unavailable protein (PC), total protein (TP), and total degradable (Total RDP) and undegradable protein (Total RUP) (Table 13) of canola seeds. Residue at 0 h and the soluble fraction of canola seeds are related to peak area, amide II area, amide areas ratio, amide I and II heights, and β-sheet height (Table 15).
Theodoridou and Yu [5,6] studied the correlation of canola meals and presscake to protein structures and they also found that amide I and II areas and their ratio were related to NDIP (r = 0.95, p = 0.051), PB1 (r = -0.76, p = 0.244), PB2 (r = 0.82, p = 0.188), IDP (r = 0.89, p = 0.107), MP (r = 0.99, p = 0.006). The high r values that they obtained show a tendency for a strong relationship, but the high P values indicate that a higher sample size is necessary to confirm those relationships [16]. Similar to our results, Huang [17] found relationships between the amide I and II areas, heights and their ratios and    CP, soluble crude protein, PA2, PB1, tdCP, S, D, and TDP in pelleted canola meals (r>0.76 and p<0.05), but did not find correlation between α-helix height, β-sheet height, and α-helix: β-sheet ratio and any protein parameters of canola meals. These results, along with ours, indicate that the various processes for oil extraction, desolventizing of the meals, and pelleting may affect the protein structures of the meals differently, even if the companies use similar processes. Or simply, they indicate that repeating the study with higher sample sizes would improve the results and give us a clearer understanding of the correlations between protein structures and the characteristics of canola meals. Although based on our results it seems to be easier to relate protein spectral structures   with the protein structures of canola seeds, because more frequent and stronger relationships could be observed for the seeds than for the meals, more repetition could only help to support the results presented here.

Prediction of nutrient supply and rumen and intestinal digestion of canola seeds and meals using unique molecular spectral features
Multiple regression analysis is used to verify the strength of the relationship between a dependent variable and several predictor variables, quantifying and statistically eliminating the effect of other predictors [18]. In our study, carbohydrate and protein-related spectral structures were used to predict chemical, degradable, and digestible characteristics of canola meals and seeds. However, the most significant prediction equations were relating protein-related structures to ruminal degradability and intestinal digestibility aspects of canola seeds (Table 20). The area of Amide I along with either the heights of α-helix or β-sheet or their ratio seem to be good predictors of rumen degradable and undegradable soluble and moderately soluble protein fractions (PA2, RDPA2, RDPB1, RUPA2, and RUPB1), and of the total rumen undegraded protein (Total RUP) in canola seeds (p<0.05, R 2 ≥0.65). These results are important because in ruminants there is an extensive use of nitrogen compounds by the ruminal microbiota and that affects the quantity and quality of protein available for digestion in the small intestine. Being able to predict with more confidence how much of the protein in the canola seeds will be available for the animal to use, is extremely helpful in animal nutrition. Although a low R 2 means variation in the results, this variability shows a trend behavior of the variables studied. All the prediction equations for canola meals using proteinrelated structures showed low R 2 but they also showed very low p-values (Table 20). This indicates that even not being too precise, a trend is observed between those variables, and α-helix height is the protein-related structure that appears to be a good predictor for many energy-related variables. Crude protein and total digestible crude protein (TDCP) also showed a trend to be predicted by the α-helix height and amide II height. A similar response can be observed between some carbohydrate-related structures and some aspects of canola meals and seeds (Table 21). These results clearly show a pattern and further analysis with more data would likely increase the R 2 values and give more assurance to the users of these equations.

CONCLUSION
The correlation study between carbohydrate spectral features and canola meals and seeds showed that STCA commonly appear to be related to canola meals and TCA to canola seeds features. And the correlation between protein spectral features and canola meals and seeds aspects showed strong relationships with the amide region of both seeds and meals, but more and stronger relationships were observed on canola seeds. These results indicate that the carbohydrate and protein structures obtained with FTIR-ATR have been proven to be related to aspects of canola seeds and meals' chemical and nutrient profiles, as well as rumen degradable and intestinal digestibility characteristics. Also, the multi-regression analysis of canola meals and seeds and carbohydrate and protein-related molecular structures showed trends between protein-related structures for the canola meals equations (p≤0.004 and R 2 ≥0.23). However, high R 2 (>0.64) and low p

CONFLICT OF INTEREST
We certify that there is no conflict of interest with any financial organization regarding the material discussed in the manuscript.

FUNDING
The SRP Chair (PY) research programs are financially supported by grants from the Ministry of Agriculture Strategic