Challenges in inbreeding estimation of large populations based on Polish Holstein-Friesian cattle pedigree
The aim of this study was to evaluate observed and future inbreeding level in Polish Holstein-Friesian cattle population. In total, over 9.8 mln animals were used in the analysis coming from the pedigree of Polish Federation of Cattle Breeders and Dairy Farmers. Inbreeding level, as an average per birth year, was estimated with the method accounting for missing parent information with the assumption of year 1950 as the base year of the population. If an animal had no ancestral records, an average inbreeding level from its birth year was assigned. Twice the average inbreeding level served as relatedness of the animal to the population, which enabled estimation of inbreeding in its offspring. The future inbreeding of potential offspring was estimated as an average of animals (bulls and cows) available for mating in a certain year. It was observed that 30–50% of animals born between 1985 and 2015 had no relevant ancestral information, which is caused by a high number of new animals and/or entire farms entering the national milk recordings. For the year 2015, the observed inbreeding level was 3.30%, which was more than twice the inbreeding with the classical approach (without missing parent information) and higher by 0.4% than the future inbreeding. The average increase of inbreeding in years 2010–2015 was 0.10%, which is similar to other countries monitored by World Holstein-Friesian Federation. However, the values might be underestimated due to low pedigree completeness. The estimates of future inbreeding suggested that observed inbreeding could be even lower and also increase slower, which indicates a constant need to monitor rate of increase in inbreeding over time. The most important aspect of presented results is the necessity to advise individual farmers to keep precise recordings of the matings on their farm in order to improve the pedigree completeness of Polish Holstein-Friesian and to use suitable mating programs to avoid too rapid growth of inbreeding.
KeywordsDairy cattle Pedigree completeness Population structure Relatedness
Despite the worldwide character of Holstein-Friesian cattle population, its inbreeding level is, reportedly, continuously increasing. It is partially caused by progressing globalization of dairy cattle breeding, which leads to using a similar gene pull across all the populations. Thus, the control of inbreeding level in Holstein-Friesian cattle populations is currently one of the main focuses of World Holstein-Friesian Federation (WHFF; Feddersen et al. 2016). Increase in inbreeding level leads to occurrence of inbreeding depression (Rokouei et al. 2010; Lal et al. 2013) and a decrease in genetic variation among the animals (VanRaden 2005). The presence of inbreeding depression is associated with decrease in vitality and higher susceptibility to diseases and unfavorable environmental conditions. As a consequence, a lower reproductive (e.g., low insemination success, miscarriages) and production performance as well as higher rate of genetic disorders and lethal genes are observed. Whereas the decrease in genetic variation leads to slowing down or even completely stopping of the genetic progress of traits under selection, which is a consequence of increasing homozygosity level leading also to higher rate of a lethal diseases. Increase of inbreeding level in the population of a country or mating highly related animals within one herd will cause actual economic losses. Therefore, the WHFF promotes the control of rate of increase in inbreeding over time in worldwide populations of Holstein-Friesian cattle (Feddersen et al. 2016) and not exceeding the increase of 1% (Falconer and Mackay 1996; Lynch and Walsh 1998) per generation (~4 years).
Poland is one of the countries where monitoring of inbreeding is currently being introduced. Currently, in Poland, around 36% of Holstein-Friesian cows are under genetic evaluation, but this number increases each year with new animals entering the national milk recordings (“The results of milk recordings” 2016; http://pfhb.pl). The pedigree information of the newly added cows is scarce. Therefore, to be able to estimate the actual inbreeding level in this population, it is necessary to use a method allowing to account for information from the “missing ancestors” as the traditional approach of Wright (1922) is no longer enough. Such methodology was developed by VanRaden (1992), who proposed to assign average inbreeding level from the year of birth of individuals if their pedigree had no ancestors or only one parental line available. This “assigned inbreeding level” is later used to estimate the inbreeding of the offspring. Since the inbreeding level of the animal is equal half of its additive relationship to the population, it allows estimation of the relatedness of an animal to the population. Another important aspect that can serve as a tool in monitoring the inbreeding level in the population is estimation of a future inbreeding (VanRaden and Smith 1999). The future inbreeding represents the expected level of inbreeding in the population, based on inbreeding of animals currently available for matings. It indicates the difference between the observed inbreeding level and the inbreeding that would be present without preferences for a certain animals.
Thus, the objective of this study was to estimate the observed and future inbreeding level in Polish Holstein-Friesian cattle population.
Materials and methods
Data control of the pedigree of bulls and cows with the number of animals kept for inbreeding evaluation in Polish Holstein-Friesian population
Number of excluded animals
Final number of animals
Removing a parent if offspring younger by less than 3 years
Removing dams with more than 14 calves
Removing animals that are present both as dam and sire and sex cannot be verified
Removing animals with incorrect IDs (less than six digits without country code or less than seven digits with country code)
All animals from the corrected pedigree had also information on the year of birth, which was needed for the further calculations of inbreeding level per year. If the data provided by the Polish Federation of Cattle Breeders of Dairy Farmers did not contained the year of birth of the animal, it was calculated using the assumption that the animal had to be 3 years older than the youngest offspring present in the dataset (VanRaden 1992; Aguilar and Misztal 2008).
Establishing the base year
The study of VanRaden (1992) suggested to use the base year when calculating the inbreeding level with accounting for missing parent information as the values of calculated inbreeding coefficients depended on the pedigree depth. This was done in order to have better control of the animals included in the evaluation and to avoid overestimation of inbreeding level per birth year, when only a few animals had the pedigree available in early years of the start of population. In case of Polish population, it was decided to use animals born not earlier than in 1950 as from this year the number of Holstein-Friesian in Poland cattle began to increase intensively and historically the status of the population became more stable after Second World War (Jasiorowski et al. 1988).
Animals born before 1950 were kept in dataset if they had at least two offspring born after 1950;
Parents born before 1950 were treated as unrelated and not inbred;
Animals born as a first generation after 1950 were treated as related but not inbred;
Animals born in a second and later generations had increasing level of relatedness and inbreeding.
Calculating pedigree completeness
The value of PCI-5 for each animal could have the value from 0 to 1, depending on the number of “full generations,” where maximum value of 1 means that the animal had five full generations recorded in the pedigree. It should be noted that the first value larger than zero was PCI-5 = 0.125, which meant that the animal had at least two parents and two grandparents (each from different parent). The value of PCI-5 = 0 was, thus, assigned to all animals with less than 1.5 of full generation in pedigree, and accounted for more cases than purely “animals without parents.”
Estimation of observed inbreeding
Secondly, the obtained average inbreeding was assigned to all animals from a certain birth year that did not have parents or only mother or father were present in the dataset. Finaly, the assigned values were used as a rate of relatedness of the individual to the population. It was assumed that twice the inbreeding of the animals with missing pedigree was expressing its relationship to the population. This was later used for more accurate estimations of inbreeding level in offspring (VanRaden 1992).
Estimation of future inbreeding
- 1.Selecting the bulls:
Not older than 5 years (birth year 2011–2015);
Older than 5 years if the last use was more recent than 5 years.
- 2.Selecting the cows:
Cows not older than 6 years (birth year 2010–2014);
Pooling 200 cows per birth year—1000 cows in total for a 5-year period.
“Mating” of selected animals and estimating the average inbreeding for potential offspring per birth year.
Results and discussion
The inbreeding level in any population of livestock animals is a direct consequence of selection. It is an ultimate challenge of breeding strategies to maximize response to selection without facing the consequences of increasing inbreeding level in the population (Strabel 2001). However, the estimated inbreeding level is affected by many factors not only directly linked to selection and breeding practices (e.g., selection intensity, semen, and animal import), but also pedigree completeness level, establishing or not a base year for the population or method used for its estimation. Thus it is necessary to constantly monitor and thoroughly examine the results as the increasing level of inbreeding is causing economic losses (Strabel 2001). Following the recommendations of WHFF and comprehensive studies performed in other local populations of Holstein-Friesians, e.g., UK (Kearney et al. 2004), Denmark (Sørensen et al. 2005), Canada (Miglior and Burnside 1995), and USA (Young and Seykora 1996), the Polish population also required similar evaluation. This study was conducted to evaluate the pedigree and inbreeding level in Polish Holstein-Friesian cattle population, which was performed with a method enabling to account for missing parent information.
Pedigree data evaluation
The foundation of Polish Holstein-Friesian population can be dated back to early 1900s, yet, only from 1934, with implementation of studbook regulations, all information on milk production and pedigree started to be recorded in a uniform manner (Jasiorowski et al. 1988; Goździkiewicz, 2004). Due to losing nearly 67% of all dairy cattle during Second World War (Jasiorowski et al. 1988), for this study, a year 1950 was chosen as a base year for estimation of inbreeding level. At the beginning, the animals were mostly of a foreign origin, which was a consequence of absorptive crossing of Polish Black and White Lowland cattle (Jasiorowski et al. 1988; Goździkiewicz, 2004). This was observed especially in years 1950–1959, where over 90% of animals were imported yearly and those animals became the founders for the Polish Holstein-Friesian population. In 1970s, already several dozen Holstein-Friesian bulls were available for insemination, and from 1990s, bulls used for insemination have at least 87.5% of Holstein-Friesian blood (Goździkiewicz, 2004). Since 2010, the population in Poland has 93–99% of Polish Holstein-Friesians born yearly.
Comparison of the number of Polish and foreign Holstein-Friesian cattle with pedigree completeness index based on five generations equal to 0 (PCI-5 = 0)
Year of birth
Animals with PCI-5 = 0
As for the year 2015 for cows and 2013 for bulls, the PCI-5 has been steadily increasing, and currently, it is on average 0.62 and 0.88, for cows and bulls, respectively (Figs. 1 and 2). Nonetheless, in comparison with other countries, it is a low result. For example, in Canada, Denmark, and UK, the PCI-5 on the level of 0.90 or higher is present in the local Holstein-Friesian populations already since early 2000s (Kearney et al. 2004; Sørensen et al. 2005; Stachowicz et al. 2011). The only solution to improve the pedigree completeness in Polish Holstein-Friesian population is raising awareness among farmers and breeders to keep accurate recordings on the matings performed in their herd. It has to be noted that underestimation of inbreeding level of an individual leads to overestimation of its breeding value and could result in selecting animals that are too related to the population. Only complete pedigree information will enable accurate estimation of relatedness within the population and with this, a more precise control of the change in inbreeding rate over time.
Nonetheless, to assure that the VanRaden’s algorithm (1992) applied to Polish pedigree does not overestimate the inbreeding of known foreign bulls, the values obtained in the present analysis were compared with the data from Canadian Dairy Network (CDN; https://www.cdn.ca). The comparison indicated that on average the estimates of inbreeding level for bulls (born in years 1952–2014) present on CDN website were by 1.46% higher than those from the evaluation of Polish Holstein-Friesian: 4.86 and 3.39%, respectively; whereas when the base year was not applied to the Polish dataset, the estimates were on average by 1.96% higher than the value reported in the international database, reaching 6.82%. The previous study estimating inbreeding in Polish Holstein-Friesian population suggested that using the base year is removing important part of information from the pedigree (Jankowski 2007). However, in this study, it can be concluded that the VanRaden’s algorithm (1992) accounts for most of the missing information in Polish pedigree. Also, using the base year is crucial to avoid overestimation of the inbreeding of foreign animals.
Methodology vs. inbreeding estimation
Establishing the actual base year in any method remains debatable, as changing it directly affects the estimated level of inbreeding in the population, even though it does not affect the rate of increase in inbreeding over time. It was observed in the preliminary analysis that the change of rate inbreeding was not affected by the method selected nor the base year applied during the pedigree editing (results not shown). As presented in Fig. 6, also after establishing a base year and using three different methods for inbreeding estimation, its trend over time is very similar. This is also one of the WHFF’s assumptions and recommendations in the program monitoring the inbreeding in the world’s Holstein-Friesian populations, as the method and the level of inbreeding estimated with it is less important than the rate of increase in inbreeding over time (Fedderson et al. 2016). Thus, despite other countries using different methods to estimate the inbreeding levels in their Holstein-Friesian populations, e.g., Meuwissen and Luo (1992) in UK, Boichard (2002) in Dennmark, or Sargolzaei et al. (2005) in Canada, the rate of increase in inbreeding can be compared between them.
The inbreeding coefficient calculations in an ideal situation describe the actual inbreeding level in the population or of the single individual. However, when dealing with the actual pedigree data of the population, the calculations might be more or less further from the true values. Especially, in case of Polish pedigree, where the method accounting for missing parent information had to be used to recreate the relationship between the animals in the population, it is important to know that it did not affect the rate of increase in inbreeding. The only more precise method would be the estimation of the genomic inbreeding coefficient based on actual genotypic relationships between the animals (e.g., Leutenegger et al. 2003; VanRaden 2008; Bjelland et al. 2013). However, this is much more costly analysis than traditional pedigree-based inbreeding estimation and such data is not yet available for the Polish Holstein-Friesian population. Therefore, it needs to be assumed that despite its imperfections, assigning the average level of inbreeding from birth year to animals with unknown parents provides an inbreeding level closer to its actual value in the population and with appropriate trend over time.
Observed vs. future inbreeding levels in polish population
Comparison of polish population with other countries
The reason for these low values could be firstly the fact that Polish population is still importing nearly half of used semen and high number of animals from other countries. This means that foreign genetic material is constantly added to the Polish Holstein-Friesian population. Naturally, those animals or semen are not outcrossed in comparison to Polish cattle, as the globalization of Holstein-Friesian cattle is progressing, but have other pedigree lines than local population, which affects the estimation of inbreeding level. Secondly, quite low estimates for Polish population are directly linked with pedigree structure, depth, and low completeness that are most probably causing underestimation of the inbreeding level. Even use of the method allowing for accounting for missing parental information cannot recover it fully. Nonetheless, the rate of increase in inbreeding over time remains the same despite the applied method (VanRaden vs. classical approach).
Another step that has to be taken into consideration in inbreeding level control of Polish Holstein-Friesian population is the estimation of genomic inbreeding (Wiggans et al. 1995; VanRaden et al. 2011). Such analyses are performed in USA on regular bases since 2009 by Council of Dairy Cattle Breeding (www.uscdbc.com) and indicate far higher levels of genomic (~10.2%; www.uscdcb.com 2016) than pedigree-based inbreeding (~7.7%; www.uscdcb.com 2016). To implement genomic evaluation of inbreeding level in Polish population, the sufficient number of genotypes in necessary. As the number of genotyped Holstein-Friesian cattle in Poland is low, it is crucial to collaborate with organizations such as EuroGenomics to allow implementing the new methodology together with other countries, which should be the next step in inbreeding control.
This study aimed to estimate observed and future inbreeding levels in Polish Holstein-Friesian population. This was required to enable the control of rate of increase in inbreeding over time. The obtained values indicated that the inbreeding in Polish population is on a quite low level and that its increase over the past decades remains within the value recommended by WHFF. However, the values might be underestimated due to low level of pedigree completeness. The estimates of future inbreeding suggested that the level of observed inbreeding could be even lower and also increase over time slower, which indicates the need to monitor rate of increase in inbreeding over time. The most important aspect of the presented results is the necessity to advise individual farmers to keep precise recordings of the matings on their farm in order to improve the pedigree completeness of Polish Holstein-Friesian and to use suitable mating programs to avoid too rapid growth of inbreeding. For the future, an application of genomic data to estimate genome-based inbreeding coefficient seems to be the best option to avoid issues with pedigree quality.
ESK acknowledges financial support of the Foundation for Polish Science (FNP START 2016 grant no. 94.2016). The authors would like to thank Danuta Radzio and Piotr Goździkiewicz from Polish Federation of Cattle Breeders and Dairy Farmers for their expertise and valuable comments during pedigree data evaluation.
Compliance with ethical standards
The pedigree data used in this study were collected during routine herdbook recordings within the breeding 309 program and as such did not require an approval from Animal Care and Use Committee.
- Boichard, D (2002) Pedig: a Fortran package for pedigree analysis suited for large populations proc. 7th world Congr Genet Appl Livest Prod, Montpellier, FranceGoogle Scholar
- Falconer DS, Mackay TFC (1996) Introduction to quantitative genetics, 4th edn. Pearson Education Limited, EssexGoogle Scholar
- Feddersen E, Van Doormal B, de Jong G, Rensing S (2016) World inbreeding trend in Holsteins. World Holstein Conference, Buenos AiresGoogle Scholar
- Jankowski, T. (2007) Optimization of mating design with constrained inbreeding of Polish Holstein-Friesian. PhD Thesis. Poznan university of life sciencesGoogle Scholar
- Jasiorowski, H. A., M. Stolzman, and Z. Reklewski (1988) The international Friesian strain comparison trial. A world perspective. Food and agriculture organization of the United NationsGoogle Scholar
- Krychowski, T., and A. Nowosielska. 2016. Analiza użytkowania nasienia buhajów w pierwszym półroczu 2015 roku. 2:4–8Google Scholar
- Goździkiewicz L (ed) (2004) 100 lat oceny wartości użytkowej bydła w Polsce. Krajowe Centrum Hodowli Zwierząt, WarszawaGoogle Scholar
- Lal C, Kumar V, Maloo SR (2013) Heterosis and inbreeding depression for some quantitative and heat tolerance characters in bread wheat (Triticum aestivum L.). J Wheat Res 5:52–55Google Scholar
- Lynch M, Walsh B (1998) Genetics and analysis of quantitative traits, 1st edn. Sinaure, Sunderland, MAGoogle Scholar
- Strabel T (2001) Ograniczanie inbredu we współczesnej hodowli bydła mlecznego. Prace i Materiały Zootechniczne 59:25–36Google Scholar
- VanRaden PM (2005) Inbreeding adjustments and effect on genetic trend estimates. Interbull Bull 33:81–84Google Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.