Food purine data provenance
Source: USDA and ODS-NIH Database for the Purine Content of Foods, Release 3, September 2026. The original workbook is unchanged in source/. Source URL, retrieval date, byte count and SHA-256 are recorded in asset-downloads.json.
The extraction report confirms 501 food records from two food worksheets and 15 alcoholic-beverage records from the separate beverage worksheet. Dietary supplement worksheets remain available in the original workbook but are deliberately not mixed with foods. A supplement dose is not a food portion.
Definitions
total_purines: original cached total of adenine, guanine, hypoxanthine and xanthine. The JSON/CSV preserves source precision; the Markdown table rounds display to one decimal place.- Food unit: mg/100 g, preserving the named raw, cooked, dried or other state. Alcoholic beverages: mg/100 mL. The source landing page contains inconsistent shorthand; worksheet headers and documentation resolve the denominator.
uric_acid_measured: a separate source column; never add to total purines or confuse with an older calculated uric-acid equivalent.ND: not detected; not proven zero. A dash means not measured/calculable in source documentation, not zero. Two food totals (butter and margarine) are ND and are preserved as such.n, individual-base SEM/min/max and original source codes/year/country are retained. Total-purine uncertainty is not fabricated by summing base SEMs.source_sheet,source_row,source_total_cell: exact workbook provenance. Identically named foods from different regions or studies remain distinct.usda_analytical_bold: whether the source food-description cell is bold, which the workbook uses to identify USDA analytical data. This is a presentation-derived provenance flag, not a new quality score.
Food portions may be estimated as total mg/100 g × portion grams / 100, only for the same food state and a numeric total. Do not infer the cooked value from a raw entry or the milliliters of a liquid from grams without density. No universal daily purine allowance or safe/unsafe food classification is supplied.
Limits and maintenance
Values are sample estimates, not exact composition of every brand, harvest or recipe. Some legacy literature has one sample; cooking and sample selection vary. Missing foods remain missing. The source workbook has merged headers and footnotes, so rerun the explicit extractor rather than importing the first row as generic column names. The extraction script verifies expected headers and base-total arithmetic where all components are numeric. Original numeric totals and all ND markers must round-trip to the exact source cells.
The study-derived composition data do not establish clinical outcomes for those foods. Plant and animal purine exposures differ in observed associations; alcohol and sugar can matter even when purines are low. Refer to clinical food pages for interpretation. Original-source attribution should remain visible when data are used. Do not republish source-paper prose or figures merely because numerical data are accessible.