Estimating Animal and Plant Protein Intakes in Diet Assessed by Automated Self-Administered 24-h Recall (ASA24) and Food Records Authors: Tuo Lan, Yikyung Park, Natasha Tasevska Published: 2026 ABSTRACT AbstractBackgroundDespite growing interest in the distinct roles of animal protein (AP) and plant protein (PP) in health, commonly used dietary assessment instruments estimate only total protein intake and do not differentiate between protein sources.ObjectivesThis study aimed to develop a protocol to estimate AP and PP intake in diets assessed by the automated self-administered 24-h recall (ASA24) and food records.MethodsWe used data from the Interactive Diet and Activity Tracking Study, including 926 participants (mean age, 63.1 y). Participants completed ≤6 ASA24s and two 4-d food records (4DFRs) over 12 mo. AP and PP were defined as proteins derived from animal- and plant-based foods, respectively. AP and PP were estimated by linking foods reported on ASA24 and 4DFR to the USDA Food and Nutrient Database for Dietary Studies, USDA National Nutrient Database for Standard Reference, MyPyramid Equivalents Database, and Nutrition Data System for Research. The validity of the method was assessed by comparing the original total protein intake to the summed AP and PP estimates.ResultsA total of 3171 and 3652 unique foods were reported on the ASA24s and 4DFRs, respectively. The mean difference between the original total protein values and the sum of AP and PP was negligible for both ASA24 (0.027 g/100 g food; SD = 0.63) and 4DFR (0.019 g/100 g food; SD = 0.69). Approximately 89% of ASA24 and 88% of 4DFR foods had differences within ±0.1 g/100 g. On the basis of ASA24, AP accounted for 61% of total protein intake in males and 57% in females. The top sources of AP were red meat (19.3%), poultry (14.0%), and cheese (10.4%). In contrast, PP was mainly from bread (19.6%), nuts and seeds (10.9%), and pasta (7.1%).ConclusionsThis protocol provides a practical tool for estimating AP and PP intake from ASA24, thereby advancing nutritional epidemiologic research and supporting the development of evidence-based dietary guidelines.Keywords: animal protein, plant protein, protein, protein source, ASA24, food record, dietary assessment Introduction Consumption of a variety of high-quality nutrient-dense protein sources is recommended for supporting overall health, reducing the risk of chronic disease, and meeting nutrient needs [[1], [2], [3]]. Although more than half of the United States population meets the overall protein intake recommendations, many do not meet the recommendations for specific protein subgroups, such as nuts, seeds, soy products, and seafood [3]. Animal-based protein sources, such as meat and dairy, provide higher levels and appropriate proportions of indispensable amino acids relative to human requirements, yet contribute to a higher saturated fat, sodium, and a lower dietary fiber intake, which are established risk factors for cardiometabolic diseases [[4], [5], [6]]. Furthermore, there is some evidence linking branched-chain amino acids (leucine, isoleucine, and valine), indispensable amino acids particularly high in animal-based proteins, to insulin resistance and shortened life span [[7], [8], [9]]. In contrast, plant-based proteins from vegetables, legumes, nuts, and seeds generally have a higher level of arginine and glycine, which, along with other nutrients present in plant foods, such as carotenoids, fiber, and bioactive compounds, support anti-inflammatory pathways and reduce oxidative stress [10,11]. High intake of animal protein (AP), particularly from processed red meats, has been consistently associated with an increased risk of diabetes, all-cause cancer, cardiovascular disease, and other cause-specific mortality, whereas intake of plant protein (PP) has been shown to lower all-cause and cardiovascular disease mortality risks [[12], [13], [14], [15]]. Despite growing interest in the health effects of AP compared with PP sources, a limited number of studies have been able to evaluate source-specific protein intake due to limitations in commonly used dietary assessment instruments [16]. The automated self-administered 24-h recall (ASA24), a widely used web-based tool developed by the National Cancer Institute for large population-based studies in the United States [12], which uses the USDA’s Food and Nutrient Database for Dietary Studies (FNDDS) to estimate nutrient intake [17,18], provides estimates of total protein intake but does not differentiate between AP and PP. Similarly, any instrument that uses the FNDDS database, which only includes information on the total protein content of foods, is affected by this limitation [19]. The Interactive Diet and Activity Tracking in American Association of Retired Persons (IDATA) study is a large validation study established by the National Cancer Institute to evaluate internet-based and traditional self-report tools for dietary and physical activity assessment in epidemiologic research [20]. In this study, we used data from the IDATA to develop a protocol to estimate AP and PP intake in the diet assessed by ASA24 and a 4-d food record (4DFR) that used the same databases for nutrient estimation [20]. Methods Study design and population The IDATA study included 1110 participants who were 50 to 74 y old at the time of the study (January 2012–September 2012). The study has been previously described in detail elsewhere [20]. Briefly, participants were recruited from the member pool of the AARP (formerly known as the American Association of Retired Persons) residing in the Pittsburgh, Pennsylvania area. Eligibility criteria included being an English-speaker, not being on a weight-loss diet, having access to the internet, and having no metabolic disorders or mobility limitations. For the present study, we used data from 926 participants who had ASA24 or 4DFR data available. All participants provided written informed consent, and the study was approved by the National Cancer Institute Special Studies Institutional Review Board. ASA24 During a 12-mo study period, participants were asked to complete an ASA24 every other month, a total of 6 ASA24s. ASA24 is an online self-administered tool that asks participants to list all foods, beverages, and dietary supplements consumed from midnight to midnight on the previous day. This was followed by detailed questions regarding food preparation methods, additions, and portion sizes. Pictures of foods helped participants estimate portion sizes. A total of 896 participants provided 4942 ASA24s. We excluded ASA24s that were incomplete (n = 52) or with total energy intake <500 kcal/d (n = 29), as such values are unlikely to reflect plausible energy intake. If a participant completed multiple ASA24s and 1 had an energy intake <500 kcal, only that specific recall was excluded, whereas the participant was still included in the analysis. As a result, 896 participants with a total of 4861 ASA24s were included in the ASA24 analysis (66.7% completed all 6 ASA24s). Nutrients from the ASA24-2011 were estimated using the USDA’s FNDDS, version 4.1 [21] and USDA National Nutrient Database for Standard Reference (SR), Release 22 [22]; MyPyramid equivalents were estimated using MyPyramid Equivalents Database (MPED), version 2.0 [23]; and supplements intake using the NHANES Dietary Supplement Database 2007–2008 [24]. 4DFR Participants were asked to complete 2 unweighted paper-and-pencil 4DFRs, administered 6 mo apart. Each record captured all foods and beverages consumed over 4 consecutive days. Participants received detailed instructions and a serving size booklet to help with reporting. Dietary supplement intakes were not reported. For both the first and second 4DFR, 832 participants had available data. We excluded 4DFRs that were incomplete, defined as records that recorded diet over <4 d (n = 41 for the first 4DFR; n = 146 for the second 4DFR) or with total energy intake <500 kcal/d (n = 10 for the first 4DFR; n = 10 for the second 4DFR). As a result, 819 participants completed a total of 1457 4DFRs (77.9% participants completed both 4DFRs, 17.4% completed only the first 4DFR, and 4.6% only the second 4DFR). Nutrients from the 4DFR were estimated using the FNDDS, version 5.0 [25] and SR, Release 24 [26]; and MyPyramid equivalents intake were estimated using MPED, version 2.0 [23]. Estimation of AP and PP We used 4 nutrient databases, including the FNDDS (versions 4.1 and 5.0), SR (releases 22 and 24), MPED (versions 2.0), and the Nutrition Data System for Research (NDSR; version 2015) (Nutrition Coordinating Center) to develop a protocol for estimating AP and PP intakes in ASA24 and 4FDR (Figure 1 and Supplementary Figure 1). We here describe the process using ASA24 as an example. Estimating animal protein and plant protein intake from ASA24 in the IDATA study. Protein content was adjusted to account for changes in moisture and fat using the FNDDS 4.1 moisture and fat adjustment file. For unmatched SR codes (n = 10), protein values were obtained from the FNDDS 4.1 database. AP, animal protein; ASA24, automated self-administered 24-h recall; FNDDS, Food and Nutrient Database for Dietary Studies; IDATA, Interactive Diet and Activity Tracking in AARP; MPED, MyPyramid Equivalents Database; NDSR, Nutrition Data System for Research; PP, plant protein; SR, Standard Reference (USDA National Nutrient Database for Standard Reference); SR22, USDA National Nutrient Database for Standard Reference, Release 22. A total of 3171 unique foods were reported on the ASA24s, identified by an 8-digit FNDDS food code (Figure 1). First, the food codes were linked to their SR ingredient codes via the FNDDS 4.1-SR22 links file. Based on the number of SR ingredient codes linked to each of the 3171 unique FNDDS food codes, the FNDDS food codes were classified as either: 1) single-ingredient foods, defined as foods linked to a single SR ingredient (n = 604), or 2) multiple-ingredient foods that linked with >1 ingredient (i.e., food mixtures) (n = 2567). For the 604 single-ingredient foods identified from the ASA24, we used the MPED output to initially assign the protein source as animal or plant. The assigned sources were then independently verified by 3 investigators (TL, YP, and NT) using the SR code descriptions and reassigned, if needed. For example, at the initial classification step, soy milk (FNDDS code = 11320200) got linked to the MPED dairy group, as this is how MPED classifies soy milk, and was thus incorrectly assigned as AP. However, after the verification using the SR code description “Soymilk (all flavors), nonfat, with added calcium, vitamins A and D” (SR code = 16230), it was reclassified as a PP. Total protein values were obtained from the SR22 database, where available; for unmatched SR codes (n = 10), the FNDDS 4.1 database was used to obtain the total protein values. Finally, the protein content was adjusted for moisture and fat changes that occur during food preparation using factors from the FNDDS 4.1 Moisture and Fat adjustments file (Figure 1 and Table 1). TABLE 1Examples of estimated animal protein and plant protein contents in foods in ASA24 in the IDATA study.TABLE 1ClassificationsFNDDSSR ingredientCalculated values per SR ingredientTPAPPPCodeDescriptionMoisture change (%)1Fat change(%)1CodeDescriptionWeight(g)(g/100 g SR)A. Single ingredient11320200Milk, soy, nonfat, ready-to-drink, not for babies0016230Soymilk (all flavors), nonfat, with added calcium, vitamins A and D1002.4702.47B. Multiple ingredients a) Single protein source SR27510210Cheeseburger, plain, on a bun0018350Rolls, hamburger or hotdog, plain529.509.513317Beef, ground, patties, frozen, cooked, broiled3823.0523.101046Cheese food, pasteurized process, American, without di-sodium phosphate1318.418.402047Salt, table0.2000Total103.215.5910.814.79 b) Mixed protein source SR, matched with NDSR27550000Fish sandwich, on a bun, with spread–9.23.715028Fish, flatfish (flounder and sole species), raw57.618.8418.8401123Egg, whole, raw, fresh3.212.5712.5701151Milk, nonfat, fluid, without added vitamin A and vitamin D (fat-free or skim)13.373.37018079Bread crumbs, dry, grated, plain6.413.35013.352047Salt, table0.400018350Rolls, hamburger or hotdog, plain659.509.581302050Tartar sauce190.6280.2430.385Total152.612.837.894.94 c) Mixed protein source SR, unmatched with NDSR53120100Cake, white, standard-type mix (egg whites and water added to mix), without icing–18018137Cake, white, dry mix, regular5674.52.252.2514429Water, tap, municipal2670001124Egg, white, raw, fresh6610.910.90Total899.634.432.701.73Abbreviations: AP, animal protein; ASA24, automated self-administered 24-h recall; FNDDS, Food and Nutrient Database for Dietary Studies; IDATA, Interactive Diet and Activity Tracking in AARP; NDSR, Nutrition Data System for Research; PP, plant protein; SR, Standard Reference (USDA National Nutrient Database for Standard Reference); TP, total protein.1Moisture and fat changes during food preparation for ASA24 were based on FNDDS 4.1 reference tables. Examples of estimated animal protein and plant protein contents in foods in ASA24 in the IDATA study. Abbreviations: AP, animal protein; ASA24, automated self-administered 24-h recall; FNDDS, Food and Nutrient Database for Dietary Studies; IDATA, Interactive Diet and Activity Tracking in AARP; NDSR, Nutrition Data System for Research; PP, plant protein; SR, Standard Reference (USDA National Nutrient Database for Standard Reference); TP, total protein. Moisture and fat changes during food preparation for ASA24 were based on FNDDS 4.1 reference tables. The FNDDS food codes linked with multiple SR ingredient codes, each to form a recipe, were identified as multiple-ingredient foods or food mixtures (n = 2567). A total of 1763 unique SR ingredient codes were part of the multiple-ingredient foods and were classified as either a single source (animal-based or plant-based) (n = 1493) or a mixed-source SR ingredients (n = 270) containing both animal- and plant-based proteins (Figure 1), based on the SR code description. For single-source SR ingredient codes, protein values were obtained from the SR22 database and were assigned as AP or PP. For mixed-source SR ingredients, the 3 investigators independently reviewed each item and assigned corresponding NDSR food codes. Then, AP and PP values were obtained from the matched NDSR food codes, where available; unmatched mixed-source SR codes (n = 11), such as cake powder, were assigned a protein composition of 50% animal and 50% PP. To obtain AP and PP levels in the multiple-ingredient foods, first, protein amounts were calculated at the SR ingredient level by multiplying the SR ingredient weight from the recipe by protein content (g/100 g SR). Second, ingredient-level AP and PP values were summed across all ingredients within each food. Third, AP and PP contents per food (g/100 g FNDDS) were calculated by dividing the aggregated protein content by the total food weight. Lastly, moisture and fat change factors were applied. For example, a fish sandwich (FNDDS Code = 27550000) is a multiple-ingredient food that is linked to a total of 7 SR ingredients, including 6 single-source SR ingredients (e.g., fish, egg, and breadcrumbs) and 1 mixed-source SR ingredient (tartar sauce) (Table 1). AP and PP values for the 6 single-source SR ingredients were obtained from the SR22 database, whereas the AP and PP values for the mixed-source SR ingredient were obtained from the NDSR database (Figure 1 and Table 1). Specifically, using the values shown in Table 1, the AP and PP contents for this multiple-ingredient food (g/100 g FNDDS) were calculated as follows: A total of 3652 unique foods were reported on the 4DFRs. Animal and PP intakes were calculated following the same protocol using the FNDDS 5.0, SR 24, MPED ver 2.0, and NDSR ver 2015 (Supplementary Figure 1). Statistical analysis To evaluate our protocol for estimating AP and PP protein, for each food, we quantified the difference between the original total protein intake in the IDATA dataset and the summed AP and PP intake values estimated using our protocol. We then examined the empirical distribution of these differences using histograms. Baseline characteristics of participants in the IDATA study were presented for all participants and separately for males and females. Continuous variables, including age, are presented as means with SD, whereas categorical variables are presented as percentages. Geometric means and their 95% confidence intervals (CIs) were calculated using data from all ASA24s and the 4DFRs. For ASA24, the presented geometric means were based on all available recalls per participant (n = 896 participants; 4861 total ASA24 recalls). For 4DFR, AP and PP intakes were estimated separately for each 4DFR and then mean across the 2 records (n = 819 participants; 1457 4DFR-1 or 4DFR-2). To evaluate the main food source of AP and PP in the IDATA study, we categorized the 3171 unique foods from the ASA24s into 43 main food groups based on the FNDDS codes and their descriptions. All statistical analyses were performed using SAS software (version 9.4; SAS Institute Inc.). Results The IDATA study population had a mean age of 63.1 y (SD = 5.9) and included almost an equal proportion of males and females (males = 49%; females = 51%) (Table 2). More than half of the participants were in the age group 60–69 y (55%), and most were non-Hispanic Whites (92%). Males were more likely to be categorized as overweight (51.0% in males and 33.6% in females), whereas the prevalence of obesity was similar between males (30.3%) and females (33.1%). Most participants (97.6% of males and 96.0% of females) completed at least one ASA24, and 84% completed at least 4 or more ASA24s. Males tended to have slightly higher completion rates than females. The completion rates for the 4DFR were lower for the second administration compared with the first. TABLE 2Characteristics of study participants in the IDATA study.TABLE 2AllMalesFemalesn = 926n = 455n = 471Age, mean (SD)63.1 (5.9)64.1 (5.6)62.2 (6.0)Age (y), % 50–5929.222.235.9 60–6954.857.652.0 ≥7016.120.212.1White, non-Hispanic, %92.195.888.5BMI, % (kg/m2) <2526.118.733.3 25–<3042.151.033.6 ≥3031.830.333.1Participants who completed, n (%) ≥1 ASA24896 (96.8)444 (97.6)452 (96.0) ≥2 ASA24869 (93.8)436 (95.8)433 (91.9) ≥3 ASA24837 (90.4)424 (93.2)413 (87.7) ≥4 ASA24779 (84.1)402 (88.4)377 (80.0) ≥5 ASA24598 (64.6)327 (71.9)271 (57.5) 6 ASA24598 (64.6)327 (71.9)271 (57.5) 4DFR-1781 (84.3)397 (87.3)384 (81.5) 4DFR-2676 (73.0)346 (76.0)340 (72.2)Abbreviations: 4DFR, 4-d food record; ASA24, automated self-administered 24-h recall; IDATA, Interactive Diet and Activity Tracking in AARP. Characteristics of study participants in the IDATA study. Abbreviations: 4DFR, 4-d food record; ASA24, automated self-administered 24-h recall; IDATA, Interactive Diet and Activity Tracking in AARP. The mean difference in total protein intake between the original protein values and the estimated sum of AP and PP was 0.027 g/100 g of food (SD = 0.63) for ASA24 and 0.019 g/100 g of food (SD = 0.69) for 4DFR. The distribution of differences between the original protein value and the estimated sum of AP and PP in the IDATA study is shown in Figure 2. For ASA24 and 4DFR, 89% and 88% of foods, respectively, had differences between the original total protein values and the estimated sum of AP and PP within ± 0.1 g/100 g. For ASA24, 0.13% of foods had estimates >5 g lower, and 0.16% had estimates >5 g higher than the original values; for 4DFR, these proportions were 0.11% and 0.19%, respectively. Distribution of differences between original total protein intake and estimated sum of animal and plant protein intakes in ASA24 and 4DFR in the IDATA study. (A) ASA24 (n = 3171 FNDDS codes). (B) 4DFR (n = 3652 FNDDS codes). 4DFR, 4-d food record; AP, animal protein; ASA24, automated self-administered 24-h recall; FNDDS, Food and Nutrient Database for Dietary Studies; MPED, MyPyramid Equivalents Database; NDSR, Nutrition Data System for Research; PP, plant protein; SR22, USDA National Nutrient Database for Standard Reference, Release 22. Males had higher total protein, AP, and PP intakes than females (Table 3). For ASA24s, the mean AP intake was 54.7 g/d (95% CI: 52.8, 56.8) in males and 40.9 g/d (95% CI: 39.0, 42.8) in females, and represented 61% and 57% of total protein in males and females, respectively. The mean PP intake was 27.3 g/d (95% CI: 26.8, 27.8) in males and 22.6 g/d (95% CI: 22.1, 23.1) in females. The sum of AP and PP animal and PP was closely aligned with the original total protein values generated from ASA24 for both sexes (males: 89.4 g/d compared with 90.0 g/d; females: 71.7 g/d compared with 71.4 g/d). Similar patterns were observed for 4DFR. Mean AP intake by 4DFR was 55.2 g/d (95% CI: 54.4, 55.9) or 63.1% of total protein in males, and 41.9 g/d (95% CI: 41.3, 42.6) or 61.3% of total protein in females. Mean PP intake was 30.2 g/d (95% CI: 29.8, 30.5) in males and 24.3 g/d (95% CI: 24.1, 24.6) in females. TABLE 3Geometric mean of animal and plant protein intake in ASA24 and 4DFR in the IDATA study.TABLE 3g/d; mean (95% CI)AllMalesFemalesASA241 Animal protein47.4 (46.0, 48.9)54.7 (52.8, 56.8)40.9 (39.0, 42.8) Plant protein24.9 (24.5, 25.3)27.3 (26.8, 27.8)22.6 (22.1, 23.1) Estimated total protein, sum of AP and PP79.9 (78.8, 80.9)89.4 (87.8, 91.0)71.7 (69.8, 72.4) Original total protein from ASA2480.3 (79.3, 81.3)90.0 (88.4, 91.6)71.4 (70.1, 72.7)4DFR2 Animal protein48.2 (47.7, 48.8)55.2 (54.4, 55.9)41.9 (41.3, 42.6) Plant protein27.2 (26.9, 27.4)30.2 (29.8, 30.5)24.3 (24.1, 24.6) Estimated total protein, sum of AP and PP77.6 (77.0, 78.1)87.5(86.7, 88.2)68.4 (67.9, 69.0) Original total protein from 4DFR77.7 (77.2, 78.3)87.6 (86.8, 88.4)68.6 (68.0, 69.2)Abbreviations: 4DFR, 4-d food record; AP, animal protein; ASA24, automated self-administered 24-h recall; CI, confidence interval; IDATA, Interactive Diet and Activity Tracking in AARP; PP, plant protein.1In 896 participants with a total of 4861 ASA24s.2The means of two 4DFRs in 819 participants. Geometric mean of animal and plant protein intake in ASA24 and 4DFR in the IDATA study. Abbreviations: 4DFR, 4-d food record; AP, animal protein; ASA24, automated self-administered 24-h recall; CI, confidence interval; IDATA, Interactive Diet and Activity Tracking in AARP; PP, plant protein. In 896 participants with a total of 4861 ASA24s. The means of two 4DFRs in 819 participants. Table 4 shows the AP and PP contents in selected foods (g/100 g). Some foods contain only AP, such as milk, yogurt, cheese, beef brisket, salmon, and boiled eggs, and some foods contain only PP, such as soy milk, nuts, and orange juice. Mixed dishes and processed foods, such as breaded pork tenderloin, chicken nuggets, breaded cod, pizza with pepperoni, lasagna, and hamburgers, provided both AP and PP. TABLE 4Amounts of animal and plant protein in selected foods (g/100 g food).TABLE 4Food itemAnimal proteinPlant proteinTotal proteing/100 g of foodMilk3.30.03.3Soy milk0.02.62.6Yogurt, plain, whole milk3.50.03.5Fruit yogurt3.60.13.7Cheese21.80.021.8Beef brisket25.70.025.7Pork, tenderloin, breaded, fried18.31.419.6Chicken nuggets13.92.015.9Breaded cod18.11.619.6Salmon, raw21.60.021.6Chili beef soup1.50.62.1Egg, whole, boiled12.50.012.5Nuts0.016.516.5White bread1.36.67.9Chocolate cupcake, with icing or filling2.22.24.4Pizza with pepperoni6.93.910.8Lasagna with meat7.03.010.0Orange juice, freshly squeezed0.00.70.7Broccoli salad with cauliflower, cheese, bacon bits, and dressing3.22.96.1Vegetable soup, cream of, prepared with milk3.00.73.7Hamburger9.75.515.2 Amounts of animal and plant protein in selected foods (g/100 g food). In the IDATA study population, the top 10 sources of AP on the ASA24 were red meat (19.3%), poultry (14.0%), and cheese (10.4%), followed by meat- or seafood-based mixed dishes, milk and yogurt, fish, processed meats, eggs and egg dishes, pizza, and pasta (Table 5). In contrast, PP was mainly from bread (19.6%), nuts and seeds (10.9%), pasta (7.1%), followed by legumes and soy products, breakfast cereals and granola, other vegetables, desserts and sweets, fruits and fruit juices, crackers, popcorn, and salty snacks, and potatoes. TABLE 5Main food sources1 of animal and plant protein in ASA24 in the IDATA study.TABLE 5Food groupRank%Animal protein Red meat119.3 Poultry214.0 Cheese310.4 Meat or seafood-based mixed dishes49.9 Milk, yogurt59.6 Fish66.6 Processed meats75.4 Egg, egg dishes, and omelets85.0 Pizza93.4 Pasta103.0Plant protein Bread119.6 Nuts, seeds210.9 Pasta37.1 Legumes, soy products46.1 Breakfast cereals and granola55.7 Other vegetables55.7 Desserts and sweet75.0 Fruits, fruit juices84.9 Crackers, popcorn, and salty snacks94.6 Potato104.0Abbreviations: ASA24, automated self-administered 24-h recall; IDATA, Interactive Diet and Activity Tracking in AARP.1The main food sources was selected out of 43 food groups:1) milk, yogurt; 2) cream; 3) ice cream, pudding; 4) cheese; 5) red meat; 6) poultry 7) organ meats; 8) processed meat; 9) fish; 10) soup; 11) egg, egg dishes, and omelets; 12) egg substitutes and prepared scrambled eggs; 13) legumes, soy products; 14) nuts, seeds; 15) bread; 16) biscuits, muffins; 17) desserts and sweet; 18) crackers, popcorn, and salty snacks; 19) pancakes, waffles, and breakfast breads; 20) pasta; 21) grains and noodles (e.g., oatmeal, rice, and noodles); 22) breakfast cereals and granola; 23) ethnic foods, mixed dishes (e.g., burrito, taco, tamale, and sushi); 24) pizza; 25) fruits, fruits juices; 26) fruit-based mixed dishes (e.g, fruit salad); 27) oil; 28) potato; 29) dark green vegetables; 30) yellow vegetables; 31) tomatoes, sauces, and soups; 32) other vegetables (e.g., sprouts, mushrooms, onions, pepper, and squash); 33) vegetable-based mixed dishes (e.g., eggplant dip, squash casserole); 34) butter, margarine, and creamy sauces; 35) salad dressings and mayonnaise; 36) sugars, syrups, sweets, and candies; 37) coffee, tea, and soft drinks; 38) alcoholic beverages; 39) water; 40) soy milk; 41) shellfish; 42) meat or seafood-based mixed dishes; and 43) burgers, sandwiches. Main food sources1 of animal and plant protein in ASA24 in the IDATA study. Abbreviations: ASA24, automated self-administered 24-h recall; IDATA, Interactive Diet and Activity Tracking in AARP. The main food sources was selected out of 43 food groups:1) milk, yogurt; 2) cream; 3) ice cream, pudding; 4) cheese; 5) red meat; 6) poultry 7) organ meats; 8) processed meat; 9) fish; 10) soup; 11) egg, egg dishes, and omelets; 12) egg substitutes and prepared scrambled eggs; 13) legumes, soy products; 14) nuts, seeds; 15) bread; 16) biscuits, muffins; 17) desserts and sweet; 18) crackers, popcorn, and salty snacks; 19) pancakes, waffles, and breakfast breads; 20) pasta; 21) grains and noodles (e.g., oatmeal, rice, and noodles); 22) breakfast cereals and granola; 23) ethnic foods, mixed dishes (e.g., burrito, taco, tamale, and sushi); 24) pizza; 25) fruits, fruits juices; 26) fruit-based mixed dishes (e.g, fruit salad); 27) oil; 28) potato; 29) dark green vegetables; 30) yellow vegetables; 31) tomatoes, sauces, and soups; 32) other vegetables (e.g., sprouts, mushrooms, onions, pepper, and squash); 33) vegetable-based mixed dishes (e.g., eggplant dip, squash casserole); 34) butter, margarine, and creamy sauces; 35) salad dressings and mayonnaise; 36) sugars, syrups, sweets, and candies; 37) coffee, tea, and soft drinks; 38) alcoholic beverages; 39) water; 40) soy milk; 41) shellfish; 42) meat or seafood-based mixed dishes; and 43) burgers, sandwiches. Discussion The goal of this study was to develop a systematic protocol to estimate AP and PP intake in diets assessed by the ASA24 and 4DFR. Applying our method, the mean differences between the original total protein and the estimated sum of AP and PP were negligible for both ASA24 (0.027 g/100 g) and 4DFR (0.019 g/100 g), indicating a strong internal validity of the protocol. Our findings that males had higher intakes of protein compared with females are consistent with nationally representative studies using data from the NHANES [16,27]. In addition, as expected, more than half of the total protein intake came from animal sources. In the NHANES study, protein was categorized into 3 mutually exclusive groups: animal, dairy, and plant [27]. AP and dairy protein together accounted for ∼62% of total protein intake (46% animal and 16% dairy). Similarly, in our study using ASA24 data, AP, which includes dairy protein, represented 61% of total protein intake in males and 57% in females. In both NHANES and our study, AP was mainly sourced from meat and poultry (e.g., chicken and cold cuts in NHANES; red meat and poultry in IDATA), whereas PP was largely obtained from breads, nuts and seeds, and pasta (both in NHANES and IDATA). Growing evidence suggests that the AP and PP intake have different roles in cardiometabolic disease, cancer, and mortality [[7], [8], [9], [10], [11], [12], [13], [14], [15]]. The nutrition profiles of animal- and plant-based proteins differ. Although AP can provide higher levels of indispensable amino acids, plant-based proteins have higher levels of arginine and glycine, important for nitric oxide synthesis, metabolic regulation, and promoting antioxidative defense, and are sourced from nutrient-dense foods high in dietary fiber and various bioactive compounds that are beneficial for health [[7], [8], [9], [10], [11]]. As such, a higher intake of AP has been associated with lower blood pressure, reduced low-density lipoprotein cholesterol levels, and improved insulin sensitivity [[28], [29], [30]]. Moreover, substituting PP for AP has been linked to a lower incidence of type 2 diabetes [31]. However, the USDA SR database, which provides and maintains nutrient composition data for the foods in the United States food supply through annual updates and serves as the main data source for any other commercially available databases for processing dietary data in the United States, only provides total protein estimates without distinguishing protein sources [32]. As a result, the ASA24, which uses nutrient composition data from the USDA SR database, does not currently provide AP and PP intake. ASA24 is freely available from the United States National Cancer Institute website for researchers to use and is thus widely used in observational, clinical, and interventional studies. Our protocol provides a method for researchers to estimate AP or PP intake levels as well as the main food sources of AP and PP when they use ASA24, food records, or any instruments using this database. We observed that nearly 20% of PP intake was from bread, whereas legumes and soy products—foods emphasized in the Dietary Guidelines for Americans—contributed only ∼6% of total PP intake. Given that the health benefits vary by protein source, our protocol can also help elucidate the role of AP and PP in health, better characterize population adherence to a healthy dietary pattern [3], hence informing future dietary guidelines. The strength of the study is that the protocol was developed using >3000 unique food codes, including both single-ingredient foods and complex multi-ingredient mixed dishes, which increases its reliability and generalizability. This protocol provided a detailed description of our approach, which can serve as a standard method allowing comparability across future studies of AP and PP intakes. Also, over 85% of food sources were identified using the publicly available USDA databases and largely automated through computer programming. By linking FNDDS and SR, we estimated AP and PP at the foods’ ingredient levels. In addition, the strong agreement between estimated and original total protein values supports the internal validity of the protocol. Although we described all steps taken, an alternative approach could be identifying protein sources directly at the SR level based solely on SR food descriptions. We used the stepwise classification to enhance efficiency by using the MPED database for initial source assignment. However, a retrospective examination suggests that the first step of classifying FNDDS food codes as single or multiple-ingredient foods can be omitted to simplify the process. Despite these strengths, there are several limitations. First, for certain mixed-source SR ingredients for which no direct NDSR match was available, we assumed that the protein was 50% animal and 50% plant. Although such cases were rare (n = 11), these assumptions may introduce misclassification. To obtain AP and PP content for the mixed-source SR ingredients, we used the NDSR dietary analysis software developed by the University of Minnesota, which requires a paid license [33]. The NDSR provides an extended list of nutrients and food components, including AP and PP, even though it is based on the USDA SR nutrient database. Finally, the classification and source identification process required review by multiple investigators, which is time-consuming and may limit scalability despite careful independent review. In summary, the developed protocol to estimate protein source from ASA24 and food records has a strong internal validity. Given the growing interest in protein source on health effects, this protocol provides a practical tool for advancing nutritional epidemiology research, which will contribute to developing evidence-based dietary guidelines. Author contributions The authors’ responsibilities were as follows – NT, YP: designed research; TL, NT, YP: conducted research, analyzed data, and wrote the article; TL, NT: took primary responsibility for the final content; and all authors: read and approved the final manuscript. Data availability IDATA data can be obtained via a proposal submission and approval process, and after an appropriate Data Transfer Usage Agreement has been executed (https://cdas.cancer.gov/learn/idata/instructions/). The code book and analytic code will be made available after publication upon request to the authors. Funding The study was funded by the US NIH/NCI 2U01CA197902. Declaration of generative AI and AI-assisted technologies in the writing process The authors declare that no AI or AI-assisted technologies were used in the writing of this manuscript. Conflict of interest The authors report no conflicts of interest. Footnotes Supplementary data to this article can be found online at https://doi.org/10.1016/j.cdnut.2026.109386. Supplementary data The following is the Supplementary data to this article: Multimedia component 1 References Associated Data Supplementary Materials Multimedia component 1 Data Availability Statement IDATA data can be obtained via a proposal submission and approval process, and after an appropriate Data Transfer Usage Agreement has been executed (https://cdas.cancer.gov/learn/idata/instructions/). The code book and analytic code will be made available after publication upon request to the authors.