Improved, expanded, and automated ancestor discovery

D.J. Null1, G.R. Wigans2, E.O.O. Ogwo3and P.M. VanRaden1

1USDA, Agricultural Research Service, Animal Genomics and Improvement Laboratory, Beltsville, MD 20705-2350, USA
2Council on Dairy Cattle Breeding, Bowie, MD 20716
3Eleanor Roosevelt High School, Greenbelt, MD 20770


2022 J. Dairy Sci.

ABSTRACT

Maternal grandsires (MGS) and maternal great grandsires (MGGS) can be discovered using percentages of haplotypes shared after removing paternal haplotypes in each generation. Accuracy of ancestor discovery was originally tested with 2011 data, published in 2013, and was retested here with 2021 data to account for more markers, different chips, revised imputation, faster generation intervals, and many more candidate grandsires. Accuracy was determined from a random sample of 88,995 calves whose dams, granddams (MGD), MGS, and MGGS were genotyped and already confirmed to have correct parent-progeny relationships. The genotypes and pedigrees of the dam and MGD were removed to determine how often the correct MGS and MGGS could be discovered within each breed. In all 5 breeds (Holstein, Jersey, Brown Swiss, Ayrshire, and Guernsey) about 92% of true grandsires were automatically filled correctly, about 5% of true grandsires were suggested but not filled, and < 2% of the added ancestors were incorrect. Most incorrect ancestors were the dam’s MGS instead of her sire or the MGD’s MGS instead of her sire. Counts across all test animals were 78,492 MGS and MGGS correctly added, 7,576 in 1st place but not added, 402 incorrectly added, and 2,525 other cases. The other cases were mostly where the true grandsire was in second place or was less than 2 years older than the dam and no grandsire was automatically added. Discovery was further improved by adjusting the birth year and haplotype sharing limits to accept and add more of the first-place candidates because most were correct. This automated system has already added hundreds of thousands of MGS for known dams and will add > 1.3 million more ancestors for animals with unknown dams or MGD using virtual dam IDs to connect calves to their MGS and / or MGGS. Any discovered ancestors thought to be incorrect can be set back to missing by animal owners. Full implementation is expected in 2022.

Keywords: genomics, grandsires, pedigrees, haplotypes