Protein accounted for 19% of the total energy intake of Australians in 2023(1). However information on the amino acid distribution of protein consumed in Australia is sparce, with only tryptophan fully represented in the 2011-2013 Australian Food and Nutrient (AUSNUT) Database(2). Our team is currently developing a machine learning model for predicting amino acid composition based on previous work matching foods from AUSNUT to the United States Department of Agriculture (USDA) Legacy Foods Database(3, 4). We hypothesised that international amino acid composition would be consistent between matched foods and our prediction model would successfully calculate values in proportion of directly analysed content. This study aimed to directly analyse amino acid composition and compare to values derived from our prediction modelling and existing global values from the USDA and Standard Tables of Food Composition from Japan(5). Diverse protein-containing foods were purchased from supermarkets and analysed in duplicate for quantification of 18 protein-bound amino acids via liquid hydrolysis (excluding cysteine and tryptophan). Comparative amino acid values were extracted from the prediction model and available food composition databases (FCDBs) from USDA and Japan. Amino acid content was compared for mean, standard deviation and coefficient of variation (CV) per amino acid for each food. The fourteen foods analysed were tofu, whole egg, processed cheese, Greek yoghurt, milk, quinoa, white bread, dry pasta, whey protein, prawns, salmon, beef, pork and chicken. The CV ranged from 0.2% in valine and up to 87% in glycine, both in the whey protein sample. The highest variance came from the model prediction of whey protein compared to the tested sample, however, when comparing direct analysis to existing FCBDs, the highest variance was 62% in glycine content of prawns. The lowest total variance was found in whole eggs with average CV at 6%, and chicken, milk and pork at 7%. The highest total amino acid CV was yoghurt (26%), whey protein (21%) and cheese (19%). The standard deviation of amino acid composition ranged from as low as 2 mg/100g methionine in whole cow’s milk, up to 1979 mg/100g glycine in whey protein. Although food production, processing and food composition analysis differ substantially between countries, variation in amino acid composition was low for high-protein foods such as eggs, chicken and pork. Higher variation was evident among commercially processed foods, including whey protein and cheese, likely reflecting the wide range of formulations and manufacturing practices across food producers. These findings support the imputation of international amino acid food composition data where data are not available in Australia. This work represents an important step in validation of our model-predicted amino acid database for AUSNUT, which could contribute to progression of a high-quality, validated tool for dietary assessment in research and clinical settings.