Diet formulation models for dairy cattle have improved significantly over the past years. Particularly when it comes to protein and amino acids (AA), moving away from the rudimentary crude protein and formulating diets to the grams of all essential AA (EAA). This has allowed for improved production predictions and feeding cattle more efficiently to reduce environmental excretions. The latest version of the Cornell Net Carbohydrate and Protein System (CNCPS v.7) has been updated significantly to use nitrogen (N) as its basis for requirements to precisely predict non-ammonia N (NAN) flows. However, there are biases in the model that lead to an over-prediction of the individual EAA flows. One of the potential solutions to this problem is updating the AA profiles of the various substrates used or described by the model. The traditional methods used for AA analysis have some shortcomings from an analytical perspective, such as the use of a single analyte internal standard, reagent instability and interference, and use of a single time point hydrolysis. Therefore, the first two objectives of this dissertation were: 1) to develop and validate a method for AA hydrolysis and analysis using zwitterion hydrophilic interaction liquid chromatography (Z-HILIC) with tandem triple quadrupole (TQ) mass spectrometry (MS) for substrates used to develop a nutrition model, and 2) to improve AA recoveries and profiles by conducting multiple time hydrolysis (MTH) with non-linear regression on the same substrates and determining correction factors (CF) for the single time point hydrolysis procedure. The Z-HILIC TQMS method was demonstrated to be an efficient, precise, and accurate method for routine analysis of AA on ruminant milk, tissue, and feeds, which does not require derivatization and improves retention of polar AA. For all substrates and using a standard reference material (SRM), the method showed standard curve linearity, low limits of detection (LOD) and quantification (LOQ), and r2 above 0.995. The intra- and inter-day precision were always measured below 15% relative standard deviation (RSD) and recoveries ranged from 75% to 118%. This analytical method was further applied to deproteinized bovine plasma and similar results were observed for analysis of AA and AA metabolites. Furthermore, as shown by others, MTH improved AA recoveries, especially for the branched-chain AA, tryptophan, serine, and threonine. Profiles were estimated using the non-linear regression model parameter A0, which represents the true AA composition of the samples, the N content of each AA, and the total N of the samples. The N-based profiles increased the amount of arginine, histidine, and lysine, which have the highest N content, and decreased the amount of methionine, phenylalanine, and tyrosine, that contain the lowest amount of N. Given that CNCPS v.7 was updated on a N basis, these profiles should allow for lower model bias when predicting EAA N flows. Since MTH requires significant amounts of time and high costs, CF are a feasible way to correct single time hydrolysis for improved AA recoveries. In the current work, a CF is estimated for each of the 26 substrates analyzed that include concentrates, forages, tissues, milk, and rumen microbes. Among these protein types, CF were significantly different for alanine, arginine, aspartic acid, glutamic acid, tryptophan, and valine. While the rest of the AA were not statistically different, the CF did vary amongst the samples analyzed and using a sample-specific CF could improve model predictions. It is noteworthy that while the present research analyzed all AA, including non-essential AA (NEAA), these AA have not been typically considered in diet formulation models because of the body’s ability to synthesize NEAA de-novo. Regardless, there are data, primarily in monogastric animals, suggesting that some of these NEAA may not be sufficiently synthesized depending on the animal’s physiological state, age, or the environment. Therefore, a third objective of this dissertation was to examine the production and metabolic responses of lactating dairy cattle to varying the amounts of EAA and NEAA while being fed a diet meeting metabolizable energy (ME) requirements but limited in metabolizable protein (MP). Five treatments were abomasally infused into twelve multiparous Holstein cows in a replicated 6 × 5 balanced incomplete block design. The treatments were: 1) Water to meet 90% of EAA and NEAA (90AA) provided by the diet; 2) EAA to meet 100% of EAA and 90% of NEAA (100EAA); 3) NEAA to meet 90% of EAA and 100% of NEAA (100NEAA); 4) EAA and NEAA to meet 100% of EAA and NEAA (100AA); 5) EAA and double NEAA to meet 100% of EAA and 110% of NEAA (110NEAA). The AA profiles for the infusions were determined using the data derived from MTH. During this trial a significant switch in the diet’s primary forage had to be done from conventional corn silage to brown mid-rib corn silage. The switch increased the diet’s digestibility, which allowed for better rumen function, higher energy, and microbial growth. This caused the base diet to be less MP limited than originally formulated and only numerical differences were observed for many production results. Nonetheless, significantly lower milk fatty acids (FA) were measured in the 110NEAA treatment caused primarily by a drop in preformed FA (PFFA). The PFFA are derived from diet and endogenous fat, and since all cows were fed the same diet, the drop was likely caused by lower lipolysis. Related to this drop, the cattle receiving the 110NEAA treatment showed significantly higher amounts of circulating plasma insulin. Plasma AA, measured using Z-HILIC TQMS, fluctuated as expected based on the treatment being infused. The 110NEAA treatment had similar NEAA plasma levels to the other treatments being infused NEAA, which led to the conclusion that these were being catabolized for other purposes. Regardless of treatment, all cows produced about 48 kg of energy-corrected milk with over 4.5% milk fat, alluding to the appropriate MP and ME supply for these animals to produce to their best potential. There were significant interactions that occurred between AA, FA, and hormones that need to be explored further. This study exemplified the importance of considering all AA when formulating diets for cattle. In conclusion, an efficient method using Z-HILIC with TQMS was developed and validated for routine analysis of AA of ruminant substrates like feeds. Faster and accurate analysis will help to update nutrient composition on-farm for nutritional models to work more effectively. Given the known setback of using chemical hydrolysis for AA analysis, MTH and non-linear regression is used as a valuable technique for enhancing AA recoveries and updating AA profiles used by the CNCPS. Since MTH is laborious and expensive, CF are suggested to improve accuracy in STH, which is the most widely used procedure by the industry and academics. Beyond methodological advancements, the research underscores the paramount importance of considering all AA in cattle diet formulation and identifying the complex interactions between AA, FA, and hormones. Continuous work on the CNCPS and other nutritional models allows for better nutrition practices on farm for more efficient and environmentally sustainable dairy production practices.