# Comparison of gene editing versus conventional breeding to introgress the polled allele into the tropically adapted australian beef cattle population

### Calculate the price of your order:

275 words
+
Approximate price

## Materials and Methods

Computer simulation using the program “ geneedit” ( Cole and Mueller, 2019 ) was used to compare introgression of the POLLED allele into the Australian Brahman population via conventional breeding or with the additional use of gene editing.

For $13.00$11.05/page, we’ll come up with flawless paper based on your requirements 311 professionals online

### Simulation Program

The “ geneedit” program simulates gene editing applied to a cattle population as an extension of Cole’s program (2015) to manage multiple recessives. Complete source code is available on GitHub and is in the public domain. The basic simulation procedures are identical to those described in detail by Cole (2015), with extensions added to support gene editing, permit more flexible selection of polled sires, and model a population with nucleus and multiplier herds.

First, base bull and cow populations are formed by simulating animals with varying ages by drawing true breeding values (TBV) from normal distributions and randomly sampling horned genotypes from a Bernoulli distribution with a parameter equal to its allele frequency in the population of interest. The program ensures that at least one carrier bull and cow will be present in the base population for each recessive. Also, to prevent a minor allele from being lost due to drift a constant mutation rate of 10 –5 is used, which results in “ AA” genotypes being converted to “ Aa” genotypes when calves are created. All base population animals are treated as founders and mated at random for 10 years to produce the population used for specific mating schemes.

Each round of the simulation represents 1 year of calendar time, and generations overlap. The population size increases each round of the simulation until the user defined maximum population size of bulls and cows is reached. Each year, animals are culled for reaching the user defined maximum permitted age. Also, if the population is too large after age-based culling then cows will be culled at random (involuntary culls) and bulls with the lowest TBV or other user defined culling criteria will be culled.

Any deviations from the basic simulation procedures described above, the specific mating schemes and gene editing strategy modeled in this Australian Brahman simulation study are described in detail in the section below and complete simulation parameters are listed in the Appendix Table 1 .

The program is a stochastic simulation, but individual replicates can be recreated if the same seed is used for random variate generation. When the software was designed the goal was to examine changes in allele frequencies in cattle populations under different mating strategies, not to create a comprehensive general-purpose simulator for animal breeding. Because of that focus, selection is on TBV that are assumed known without error rather than estimated breeding values (EBV). This also means that genomic selection is not supported.

### Modeling the Australian Brahman Beef Cattle Population

A population of 10 nucleus (seedstock) herds supplying breeding bulls to 200 multiplier (commercial) herds was modeled. Ten replicates of each scenario were simulated for 20 years, with overlapping generations.

All simulations were performed on a Thinkmate RAX QS6-4210 (Thinkmate, Inc., Waltham, MA, United States) workstation with four 12-core AMD Opteron 6344 processors with a clock speed of 2. 6 GHz, 512 GB of DDR3 1600 MHz RAM, and CentOS Linux EL7.

#### Estimating the Availability and Genetic Merit of Polled Australian Brahman Beef Cattle

POLLED allele frequency and the number of homozygous polled, heterozygous polled and horned sires were calculated from Australian Brahman animals with a SNP-based POLLED test (GeneSeek ®) results from 2014 to 2017 ( Lyons and Randhawa, 2020 ). It should be noted that only the genotypic data at the POLLED locus was available on these bulls. Additionally, the average EBV of the standard Australian Brahman economic selection index, Japan Ox ($JapOx), was calculated for each sire group as an indicator of overall genetic merit (Australian Brahman EBVs, 2018, unpublished data). The GeneSeek ® SNP-based POLLED test data used to derive allele frequency and genetic merit estimates in this study may not be completely representative of the Australian Brahman population because the POLLED test is bundled with parentage and higher-density SNP genotyping for the Australian Brahman single-step genomic evaluation. There may have been a non-random selection of phenotypically polled animals submitted to determine if they were homozygous polled. However, this dataset included a large number of herds and different sires, so estimates based on this dataset were used for further modeling. All values are in Australian dollars. Two-tailed, unpaired student t -tests were used to determine if the average$JapOx value for each genotype was significantly different from one another. P -values of ≤0. 05 were considered to be significantly different.

#### Northern Australia Beef Production Practices

Northern Australia is a vast region, covering approximately 400 million ha located within the subtropics and tropics. Major characteristics of this production system are low stocking rates (up to 1 animal unit per 150 ha), large management groups (500–1, 000 animals), multi-sire matings, and irregular handling and husbandry of cattle—typically only twice annually ( Bortolussi et al., 2005 ; Burns et al., 2010 ). In large part due to the extensive management system, it is estimated that AI is used by less than 1% of northern Australian breeding herds ( Meat and Livestock Australia Limited [MLA], 2015 ).

Due to the challenging tropical environment, approximately 85% of northern Australian cattle are B. indicus and B. indicus cross, primarily through the use of Brahman cattle ( Bortolussi et al., 2005 ). Several northern Australia studies, reviewed by Burns et al. (2010), have demonstrated that tropically adapted heifers do not reach puberty as yearlings, but the majority will by 24 months of age. Therefore, heifers are typically first joined to calve as 3-year-olds and re-joined annually unless culled. Calf losses to branding or weaning are considerable, ranging between 4 and 31% across cow populations of mixed age, breed and a range of production environments ( Burns et al., 2010 ).

#### Simulation Base Population Structure

In this simulation, the seedstock base population was 15, 000 cows and 600 bulls. The commercial base population was 35, 000 cows and 1, 500 bulls. These unrelated base population animals were assigned a birth year from −10 to 0 (cows) or −5 to 0 (bulls) by sampling from a uniform distribution. True breeding values ($JapOx) for each animal were determined by randomly sampling from a normal distribution, with a standard deviation (SD) of$34 for both the seedstock and commercial populations, and a mean of $34 for seedstock cows and$0 for commercial cows ( Johnston and Graser, 2009 ). Base population bulls averaged one genetic SD higher than cows, $68 and$34 for seedstock and commercial bulls, respectively. Each base population animal’s horned status was determined by randomly sampling sire and dam alleles using a HORNED allele frequency of 80% ( Lyons and Randhawa, 2020 ). Moreover, the proportion of polled bulls in the base population was set to 30% heterozygous and 2. 6% homozygous. Additionally, homozygous polled bulls in the base population averaged 0. 16 SD ($5. 44) lower$JapOx than horned bulls (Australian Brahman EBVs, 2018, unpublished data).

#### Simulation Scheme

The flow of operations in the simulation is shown schematically in Figure 2 . The different mating schemes, and when applicable, gene editing, started in year 0. Females were mated to have their first calf at age 3 and bulls were eligible for breeding at age 2. Mating via natural service was modeled so each herd used a unique portfolio of bulls, bulls within a sire portfolio were mated randomly to cows in the herd, and each bull was limited to 35 matings per year. The polled/horned genotype and source of bulls (seedstock or commercial) depended on the mating scheme used (described below). Females did not move between herds.

FIGURE 2

Schematic of the flow of operations in the simulation. 1 The order of bull ranking was based on true breeding values (TBV) of the Japan Ox selection index ($JapOx) and, depending on the mating scheme used, also by polled genotype. 2 See Figures 1 , 3 for a detailed gene editing model and flowchart, respectively. Calves were born annually in the same herd as their dams. Sex was assigned randomly to new calves with a 50: 50 sex ratio. The TBV for new calves were created by taking the parent average and adding a Mendelian sampling term (MS): TBV calf = 0. 5(TBV sire + TBV dam ) + MS, where TBV calf , TBV sire , and TBV dam are the TBV of the calf, its sire, and its dam, respectively. The Mendelian sampling term was drawn from a normal distribution with a mean of 0 and a variance of $\frac{1}{2}\left[1–\frac{1}{2}\left({f}_{\text{S}}+{f}_{\text{D}}\right)\right]{{\sigma }}_{\text{a}}^{2},$ where f S and f D are the coefficients of inbreeding of the sire and dam, respectively, and ${{\sigma }}_{a}^{2}$ is the additive genetic variance of$JapOx ( Cole, 2015 ). Inbreeding was estimated without error based on pedigree relationships ( Aguilar and Misztal, 2008 ) using the program INBUPGF90 version 1. 42 ( Aguilar and Misztal, 2012 ). All base population animals were assumed to be unrelated, which is improbable in a commercial setting as polled bulls are likely to have common lineages. To determine a new calf’s horned status, one allele was sampled at random from each parent and used to construct the progeny genotype. Recessive genotypes were simulated without error, and it was only necessary to simulate genotypes for recessive alleles because pedigrees were assumed to be free of errors. Allele frequencies were updated each year by counting alleles ( Cole, 2015 ).

In scenarios that included gene editing ( Table 1 ), it was modeled as an added step to the elite sire production system proposed by Kasinathan et al. (2015). In this system, fetal tissue from the next generation of yet-to-be born bulls is genomically screened and selected, edited, and then somatic cell nuclear transfer (SCNT) cloning is used to create embryos for embryo transfer (ET) ( Figure 1 ). In our simulation, both heterozygous polled and horned seedstock male fetuses were sorted on $JapOx and the top 1% or 10% were gene edited to be homozygous polled and then cloned to produce a live calf. The gene editing technology modeled had editing and ET success rates of 61 and 21%, respectively. In order for a gene edited calf to be born both the edit and the ET had to be successful ( Figure 3 ). The “ geneedit” program allows the user to set a maximum number of attempts for editing and ET. However, as we modeled gene editing of a fetal cell line, which allowed for continuous editing attempts and confirmation of a successful edit before SCNT cloning and ET, the editing and ET processes were repeated until a successful outcome was observed. Under this idealized system no costs or time lag to achieve an edited bull were factored into the gene editing scenarios. In reality, there would be obvious logistical and economic considerations that would need to be evaluated to determine feasibility of gene editing to achieve homozygous polled bulls. As outlined in Figure 1 , in practice the proposed gene editing system is anticipated to require an additional 3–5 months to produce a gene edited, homozygous polled bull. TABLE 1 Parameters and results of each scenario. FIGURE 3 Flowchart for gene editing and embryo transfer (ET) process in the simulation. 1 Potential calves were sorted by their true breeding values (TBV) for the Japan Ox selection index ($JapOx) (highest to lowest).

New calves were subject randomly to a pre-weaning calf loss of 8% (seedstock) or 13% (commercial). Additionally, new horned calves of both populations were subject randomly to a dehorning mortality rate of 2% ( Bunter et al., 2013 ). Removal of scurs, which are sometimes mistaken for horns and occur in some heterozygous polled animals, could also decrease calf survival rate, but was not considered in this simulation.

Seedstock bull calves were kept for 1 year and then as yearling bulls they were sorted by $JapOx ranking (highest to lowest) and, depending on the mating scheme used, also by polled genotype. The seedstock population retained the top 5% of their yearling bulls (as defined by the mating scheme) for breeding to seedstock cows and the remainder of seedstock yearling bulls were moved to the commercial population for mating to commercial cows. In contrast, commercial population bull calves were all castrated and sold for beef annually, unless noted otherwise in a mating scheme. Animals were culled at the end of a simulated year. In both populations, cows were culled first by age (≥10 years) and then in order to not exceed a maximum populations size of 3, 000 seedstock (∼1, 800 breeding age) and 100, 000 commercial (∼61, 000 breeding age) females, they were also culled through random selection, which modeled involuntary culling. Bulls were culled first by age (≥5 years, i. e., after 3 years of breeding). Additionally, to maintain the allowable population size bulls were culled by the lowest$JapOx ranking and variably by polled genotype, depending on the mating scheme used, [culling order: (1) horned, (2) heterozygous polled, (3) homozygous polled]. The maximum allowable population sizes were 60 seedstock and 2, 000 commercial mating-age-bulls (i. e., ≥2 years old).

#### Simulation Mating Schemes

Four total mating schemes, one baseline (A), two polled preference (B and C) and one obligatory polled (D) were modeled ( Table 1 ). The three polled mating schemes (B, C, and D) were modeled using conventional breeding methods alone, and the addition of gene editing the top 1 and 10% of seedstock bull claves per year, for a total of 10 scenarios.

For retention in the seedstock population, baseline scheme A ranked the seedstock yearling bulls solely on $JapOx. Whereas polled mating schemes (B, C, and D) ranked the seedstock yearling bulls first on polled genotype [retaining order: (1) homozygous polled, (2) heterozygous polled, (3) horned] and then on$JapOx. In baseline scheme (A) and polled preference schemes (B and C), all bulls used for breeding were sourced from the seedstock population.

To establish a baseline and model current practice, scheme A used $JapOx as the sole sire selection criterion (i. e., all polled/horned sire genotypes could be used for breeding). In polled preference scheme B, polled bulls (both heterozygous and homozygous) were preferentially used for breeding, regardless of their genetic merit ($JapOx). If the maximum number of matings was reached for all polled bulls before all females were covered, horned bulls were used to cover the remaining females. Polled preference scheme C preferentially used only homozygous polled bulls for breeding and if needed, both heterozygous polled and horned seedstock bulls were used to cover the remaining females.

### Seedstock Population

#### HORNED Allele Frequency

To model the current Australian Brahman population, all scenarios started with a HORNED allele frequency of 80% ( Figure 5A ). In baseline scenario A, which placed no selection pressure on polled, the seedstock population HORNED allele frequency remained near 80% throughout all 20 years. Consequently, after 20 years of baseline scenario A the seedstock population was estimated to consist of only 5% homozygous polled, 34% heterozygous polled and a majority (61%) of horned animals ( Figure 6A ). In contrast, selection of polled sires in the conventional breeding scenarios (B, C, and D) decreased the 20-year HORNED allele frequency by 69. 1% (95% CI [68%, 70%]), 69. 7% (95% CI [68%, 71%]), and 71. 4% (95% CI [70%, 73%]), respectively ( P ≤ 0. 05 compared to A; Figure 5A ).

FIGURE 5

Effect of each scenario on(A, B) HORNED allele frequency,(B, C) inbreeding, and(E, F) genetic merit for the seedstock(A, C, E) and commercial(B, D, F) populations. Conventional breeding scenarios (A, B, C, D) are solid lines, gene editing 1% scenarios (B_1%, C_1%, D_1%) are dashed lines, and gene editing 10% scenarios (B_10%, C_10%, D_10%) are dotted lines. Error bars represent SEM.

FIGURE 6

Estimated genotype frequencies in year 20 of each scenario for the seedstock(A) and commercial(B) populations. Genotypes are represented by a color scale from darkest (homozygous polled) to lightest (horned).

The 20-year HORNED allele frequencies of polled preference conventional breeding scenarios, B and C, were 8. 9% (95% CI [8. 4% 8. 9%]) and 8. 3% (95% CI [7. 8%, 8. 8%], P ≥ 0. 05 compared to B), respectively ( Figure 5B ). As a result, after 20 years both polled preference scenario B and C’s commercial populations were estimated to be ∼83% homozygous polled and only ∼1% horned ( Figure 6A ). The addition of gene editing the top 1 and 10% of seedstock bull calves per year (B_1% and B_10%; C_1% and C_10%) similarly decreased the 20-year HORNED allele frequency ( P ≥ 0. 9 compared to B and C, respectively).

The obligatory polled scheme D resulted in the lowest 20-year HORNED allele frequency of 6% (95% CI [5. 6%, 7. 1%]). In scheme D only homozygous polled sires were used, thus conventional breeding and the addition of gene editing the top 1 and 10% of seedstock bull calves per year resulted in the same ( P ≥ 0. 9 compared to D) rapid decrease in HORNED allele frequency ( Figure 5A ). Additionally, scheme D was estimated to result in the highest percentage of homozygous polled (87%) and 0 horned animals in year 20 ( Figure 6B ).

#### Inbreeding

The seedstock population inbreeding levels inbreeding remained below 1% in all scenarios ( Figure 5C ). In baseline scenario A, inbreeding reached 0. 70% (95% CI [0. 64%, 0. 76%]) in year 20. The polled selection conventional breeding scenarios, B, C, and D resulted in a similar 20-year inbreeding level of 0. 70–0. 76% (95% CI [0. 64%, 0. 82%], P ≥ 0. 05 compared to A). The addition of gene editing the top 1% of seedstock bull calves per year to the polled preference mating schemes (B_1% and C_1%) increased 20-year inbreeding to 0. 85% (95% CI [0. 79%, 0. 91%], P ≤ 0. 05 compared to A). However, the addition of gene editing in all other scenarios (B_10%, C_10%, D_1%, and D_10%) resulted in similar 20-year inbreeding levels ( P ≥ 0. 05 compared to A). The inbreeding at year 20 based on an estimation of

$\frac{1}{8{N}_{\text{m}}}+\frac{1}{8{N}_{\text{f}}}$

( N m is number of mothers, N F is number of fathers) was 0. 22% for all scenarios, which is lower than that calculated by the simulation program due to the fact that the number of parents does not consider the relatedness between the seedstock animals as our simulation did. This in turn was higher than the inbreeding of 0. 15% which would be expected under random mating.

#### Genetic Gain ($JapOx) Baseline scenario A, which placed no selection pressure on polled, resulted in a 20-year$JapOx value of $226 (95% CI [$224, $228]; Figure 5E ). Selection of polled sires in the conventional breeding scenarios (B, C, and D) decreased the 20-year$JapOx value by $32 (95% CI [$28, $37]),$28 (95% CI [$24,$33]), and $48 (95% CI [$43, $52]), respectively ( P ≤ 0. 05 compared to A). The polled preference conventional breeding scenarios B and C’s 20-year$JapOx values were $194 (95% CI [$192, $196]) and$198 (95% CI [$196,$200], P ≤ 0. 05 compared to B), respectively ( Figure 5E ). The addition of gene editing the top 1% of seedstock bull calves per year to these mating schemes (B_1% and C-1%) significantly increased the 20-year $JapOx value by$22 (95% CI [$17,$26], P ≤ 0. 05 compared to B) and $17 (95% CI [$12, $21], P ≤ 0. 05 compared to C, P ≥ 0. 05 compared to B_1%), respectively. Moreover, the addition of gene editing the top 10% of seedstock bull calves per year (B_10% and C-10%) further increased the 20-year$JapOx value by $31 (95% CI [$26, $36], P ≤ 0. 05 compared to B) and$29 (95% CI [$25,$34], P ≤ 0. 05 compared to C, P ≥ 0. 05 compared to B_10%), respectively. In fact, the 20-year $JapOx values in scenario B_10% ($225, 95% CI [$223,$227]) and C_10% ($227, 95% CI [$225, $229]) were not significantly different ( P ≥ 0. 9) from the baseline scenario A ( Figure 5E ). The obligatory polled conventional breeding scenario D slowed the rate of genetic gain the most, resulting in a$178 (95% CI [$176,$180]) 20-year $JapOx value. The addition of gene editing the top 1 and 10% of seedstock bull calves per year to this scheme (D_1% and D_10%) significantly increased the 20-year$JapOx value by $10 (95% CI [$6, $14]) and$15 (95% CI [$11,$19]), respectively ( P ≤ 0. 05 compared to D). The 20-year $JapOx value in D_10% was similar ($193, 95% CI [$191,$195], P ≥ 0. 05) to the polled preference conventional breeding scenario B.

Overall, the addition of gene editing the top 1 and 10% of seedstock bull calves per year resulted in significantly faster ( P ≤ 0. 05) rates of genetic gain compared to the conventional breeding scenarios (B, C, and D) of each polled mating scheme ( Figure 5E ).

#### Number and Genotype of Sires Used Per Year

In year one of all scenarios, there were ∼5 (6%) seedstock homozygous polled sires used for breeding out of the 60 total sires ( Figure 7A ). Baseline scenario A and polled preference schemes B and C also used ∼22 (36%) heterozygous polled and ∼35 (58%) horned seedstock sires. In contrast, due to the limited number of seedstock homozygous polled sires, ∼55 commercial homozygous polled bulls were needed to supplement the supply of seedstock bulls in year 1 of obligatory polled scheme D.

FIGURE 7

Effect of each scenario on the number of sires used by genotype in year 1, 5, 10, 15, and 20 for the seedstock(A) and commercial(B) populations. Genotypes are represented by a color scale from darkest (homozygous polled) to lightest (horned). Bulls produced via natural mating and sourced from the seedstock or commercial population are represented by solid or dotted bars, respectively. Gene edited seedstock sourced bulls are represented by diagonal hatched bars. Solid bars below the x -axis represent the deficit of bulls necessary to breed all females in the scenario.

Baseline scenario A resulted in little change in genotype availability/use of seedstock sires throughout all 20 years. In contrast, by year 5 the polled preference conventional breeding scenarios (B and C) resulted in 56 (95% CI [52, 60], P ≤ 0. 05) more homozygous polled sires used for breeding compared to A and 0 horned sires were used. Additionally, by year 10 all sires used for breeding were homozygous polled. However, the addition of gene editing the top 1 and 10% of seedstock bull calves per year to these polled preference schemes resulted in all sires used for breeding being homozygous polled by year 5. The greatest number of gene edited sires used for breeding was 15 (95% CI [14, 16]) around year 5 in scenarios B_1% and C_1%, whereas 55 (95% CI [53, 56], P ≤ 0. 05 compared to B_1% and C_1%) were used around year 10 in scenarios B_10% and C_10% ( Figure 7A ).

Interestingly, the addition of gene editing the top 1% of seedstock bull calves per year to obligatory scheme D resulted in 5 (95% CI [2, 7], P ≤ 0. 05) fewer gene edited sires being used for breeding in year 5 compared to B_1% and C_1%. Additionally, D_10% also resulted in 9 (95% CI [5, 11], P ≤ 0. 05) fewer gene edited sires being used for breeding in year 10 compared to B_10% and C_10%. Fewer gene edited sires were used in obligatory scheme D because there were more conventionally bred homozygous polled sires available due to the mating scheme design.

There was a sequential drop in the number of gene edited sires used in all of the 1 and 10% gene editing scenarios after 5 and 10 years, respectively. The majority of sires used for breeding in these scenarios were either the offspring of gene edited sires from the previous years or produced via conventional breeding of parents both carrying the POLLED allele. In fact, by year 20 in the 1 and 10% gene editing scenarios for all the polled mating schemes (B, C, and D), only 3 (95% CI [2, 4]) and 7 (95% CI [6, 8], P ≤ 0. 05) of the sires used for breeding were homozygous polled via gene editing, respectively ( Figure 7A ).

### Commercial Population

#### HORNED Allele Frequency

To model the current Australian Brahman population, all scenarios started with a HORNED allele frequency of 80% ( Figure 5B ). In baseline scenario A, which placed no selection pressure on polled, the commercial population HORNED allele frequency remained at a similar level throughout all 20 years. Consequently, after 20 years of baseline scenario A the commercial population was estimated to consist of only 5% homozygous polled, 34% heterozygous polled and a majority (61%) of horned animals ( Figure 6B ). In contrast, selection of polled sires in the conventional breeding scenarios (B, C, and D) decreased the 20-year HORNED allele frequency by 48. 6% (95% CI [47%, 50%]), 49. 1% (95% CI [48%, 50%]), and 70. 0% (95% CI [69%, 71%]), respectively ( P ≤ 0. 05 compared to A; Figure 5B ).

The polled preference conventional breeding scenarios B and C’s 20-year HORNED allele frequencies were 29. 4% (95% CI [29% 30%], P ≤ 0. 05 compared to A) and 28. 9% (95% CI [28%, 29%], P ≤ 0. 05 compared to A, P ≥ 0. 05 compared to B), respectively ( Figure 5B ). As a result, after 20 years both polled preference scenario B and C’s commercial populations were estimated to be ∼50% homozygous polled and less than 10% horned ( Figure 6B ).

The addition of gene editing the top 1% of seedstock bull calves per year (B_1% and C-1%) resulted in a similar 20-year HORNED allele frequency ( P ≥ 0. 9) compared to conventional breeding scenarios, B and C, respectively. In contrast, the addition of gene editing the top 10% of seedstock bull calves per year (B_10% and C-10%) resulted in a 1. 2% (95% CI [0. 0%, 2. 4%], P ≥ 0. 05 compared to B) and 1. 5% (95% CI [0. 3%, 2. 7%], P ≤ 0. 05 compared to C, P ≥ 0. 05 compared to B_10%) lower 20-year HORNED allele frequency, respectively ( Figure 5B ). These slightly lower 20-year HORNED allele frequencies of scenarios B_10% and C_10% were estimated to translate to approximately 2% more homozygous polled animals in year 20 ( Figure 6B ).

The obligatory polled scheme D resulted in the lowest 20-year HORNED allele frequency of 8. 0% (95% CI [7. 5%, 8. 5%]; Figure 5B ). In scheme D only homozygous polled sires were used, thus conventional breeding and the addition of gene editing the top 1 and 10% of seedstock bull calves per year resulted in the same ( P = 1) rapid change in HORNED allele frequency. Additionally, scheme D was estimated to result in the highest percentage of homozygous polled animals (84%) in year 20 ( Figure 6B ).

#### Inbreeding

The commercial population inbreeding remained below 0. 02% in all scenarios ( Figure 5D ). In baseline scenario A, inbreeding reached 0. 005% (95% CI [0. 0039%, 0. 0053%]) in year 20. All of the preferential polled scenarios, including the addition of gene editing 1 and 10% of the seedstock bull calves per year, (B, B_1%, B_10%, C, C_1%, and C_10%) resulted in similar ( P ≥ 0. 05 compared to A) 20-year inbreeding levels of 0. 005% (95% CI [0. 0039%, 0. 0063%]). All of the obligatory polled scenarios, including the addition of gene editing 1 and 10% of the seedstock bull calves per year, (D, D_1% and D_10%) resulted in a significantly higher 20-year inbreeding level of approximately 0. 012% (95% CI [0. 0105%, 0. 0124%], P ≤ 0. 05 compared to A). These values were slightly higher than the 0. 006% 20-year inbreeding value estimated from only the number of parents in the population, which in turn was similar to the inbreeding value expected under random mating.

### Number and Genotype of Sires Used Per Year

#### Years 1–9

In year 1 of all scenarios, only commercial base population bulls were available for mating because the seedstock herds had not yet started supplying bulls to the commercial herds ( Figure 7B ). Baseline scenario A and polled preference schemes B and C used ∼1, 100 commercial bulls (58% horned, 38% heterozygous polled, and 4% homozygous polled), which was a sufficient number to breed all mature commercial females in year 1. In contrast, obligatory polled scheme D only used ∼60 commercial bulls (100% homozygous polled) to breed in year 1. Consequently, there was a deficit of ∼1, 000 bulls (i. e., ∼35, 000 commercial cows were left open) in year 1 of obligatory polled scheme D.

After year 1, baseline scenario A and polled preference schemes B and C, only used bulls sourced from the seedstock population. In contrast, due to the limited number of homozygous polled sires, commercial bulls were needed to supplement the supply of seedstock bulls for all 20 years in obligatory polled scheme D. In this scheme, using a majority of commercial (∼900) and several seedstock (∼240) homozygous polled bulls, still resulted in a deficit of ∼350 bulls in year 5 ( Figure 7B ).

#### Years 10–20

After 10 years of baseline scenario A there was little change in genotype availability of sires, resulting in only 88 (95% CI [68, 108]) homozygous polled sires (4%) used for breeding compared to 1, 217 (95% CI [1, 203, 1, 230]) horned (60%). In contrast, the polled conventional breeding scenarios (B, C, and D) resulted in 612 (95% CI [566, 658]), 685 (95% CI [639, 731]), and 1, 912 (95% CI [1, 866, 1, 958]) more homozygous polled sires used for breeding, respectively ( P ≤ 0. 05 compared to A).

In year 10, the polled preference conventional breeding scenarios B and C used 700 (95% CI [680, 720], P ≤ 0. 05 compared to A) and 773 (95% CI [753, 793], P ≤ 0. 05 compared to A and B) homozygous polled sires for breeding, respectively. The addition of gene editing the top 1% of seedstock bull calves per year (B_1% and C_1%) resulted in a similar number ( P = 1 compared to B and C, respectively) of homozygous polled sires used for breeding in year 10. In contrast, the addition of gene editing the top 10% of seedstock bull calves per year (B_10% and C-10%) resulted in 107 (95% CI [60, 153], P ≤ 0. 05 compared to B, P ≥ 0. 05 compared to C) and 89 (95% CI [43, 135], P ≤ 0. 05 compared to C) more homozygous polled sires used for breeding in year 10, respectively ( Figure 7B ).

The obligatory polled scheme D clearly used the greatest number of homozygous polled sires (2000, 95% CI [1980, 2020], P ≤ 0. 05 compared to A, B, and C) in year 10 and thereafter. However, ∼38% (821, 95% CI [811, 830]) of the bulls used for breeding in year 10 of the obligatory polled conventional breeding scenario D were commercial, which on average had lower genetic merit than seedstock bulls. The addition of gene editing the top 1% of seedstock bull calves per year to this scheme (D_1%) resulted in a similar number of commercial bulls used for breeding in year 10 (806, 95% CI [796, 815], P ≥ 0. 05 compared to D). In contrast, the addition of gene editing the top 10% of seedstock bull calves per year to this scheme (D_10%) significantly decreased the number of commercial bulls used for breeding in year 10 by 71 (95% CI [49, 94], P ≤ 0. 05 compared to D).

In year 10, the polled selection scenarios that included gene editing the top 10% of seedstock bull calves per year (B_10%, C_10%, and D_10%), used 73 (95% CI [69, 77], P ≤ 0. 05 compared to B_1%, C_1%, and D_1%) gene edited seedstock bulls, which was ∼4% of the total bulls used for breeding. By year 20 of these scenarios (B_10%, C_10%, and D_10%), only ∼1% (26, 95% CI [24, 29], P ≤ 0. 05 compared to B_1%, C_1%, and D_1%) of the bulls used for breeding were gene edited seedstock bulls ( Figure 7B ).

### Number of Animals Sold for Beef Per Year

In the baseline scenario A and polled preference schemes B and C, the maximum commercial cow population size was reached in year 10. Thereafter, in these schemes there were ∼61, 000 cows bred per year (due to 3-year mating age limit), all mature females were mated, and ∼53, 000 animals (i. e., steers, age-culled males and females, and population size-culled females) were sold for beef per year ( Figure 8A ). Strikingly, in year 10 and thereafter the polled preference conventional breeding scenarios B and C resulted in significantly more animals sold for beef per year (20-year = 470 more, 95% CI [202, 762], P ≤ 0. 05) compared to baseline scenario A ( Figure 7B ).

FIGURE 8

Effect of each scenario on the total number of commercial animals sold for beef.(A) Number and category of animals sold in year 1, 5, 10, 15, and 20 for each scenario.(B) The difference in the number of animals sold in year 20 for each polled mating scenario compared to the baseline A scenario. Error bars represent SEM.

In contrast, due to the limited number of homozygous polled sires available ( Figure 7B ), the obligatory polled scheme D reached the maximum commercial cow population size approximately 5 years later than schemes A, B, and C ( Figure 8A ). Consequently, scheme D resulted in only ∼18, 000 (95% CI [18, 082, 18, 476], P ≤ 0. 05 compared to A) animals sold for beef in year 10 ( Figure 8A ). However, in year 15 and thereafter, even the obligatory polled scheme D resulted in a significantly higher number of animals sold for beef per year (20-year = 470 more, 95% CI [172, 805], P ≤ 0. 05) compared to baseline scenario A ( Figure 8B ).

The addition of gene editing the top 1 and 10% of seedstock bull calves per year to the polled mating schemes (B, C, and D) resulted in similar ( P ≥ 0. 3) numbers of total animals sold for beef per year for each respective polled mating scheme ( Figure 8A ). Overall, all polled mating scenarios resulted in significantly more ( P ≤ 0. 05) animals sold for beef in year 20 than baseline scenario A ( Figure 8B ). This difference was due to decreased dehorning-related calf loss as a direct outcome of producing fewer horned calves.

## Discussion

Animal Health Australia (AHA) held an open public consultation and found that issues related to “ pain relief for surgical procedures” (e. g., dehorning) were the most controversial for the Australian public ( Animal Health Australia [AHA], 2014b ). All animal welfare and animal rights groups, and some academic groups pushed to mandate pain relief irrespective of the age of the animal, while many producer groups (including major national and northern Australian cattle producer groups) advised that mandating pain relief at any age is impractical. Therefore, there is an urgent need to reduce the need for dehorning. Additionally, dehorning in northern Australia has considerable economic costs for producers due to the ∼2% mortality rate associated with the process ( Bunter et al., 2013 ). Fortunately, increasing polled, or hornless, genetics presents an opportunity to address this major animal welfare concern while eliminating a major post-weaning mortality factor. However, there are trade-offs to consider.

### Current Polled Status of Australian Brahmans

We found that the POLLED allele frequency is much higher in the Australian Brahman population (20%, Figure 4 ) compared to Spurlock et al., 2014 ; Mueller et al., 2019 ). This supports the finding of Lyons and Randhawa (2020). Interestingly, the average genetic merit of Australian Brahman horned bulls was similar compared to heterozygous polled bulls. In contrast, the average genetic merit of Australian Brahman homozygous polled bulls was significantly less (AUD $6$JapOx) than both horned and heterozygous polled bulls. However, this difference is much less drastic than between United States dairy homozygous polled and horned bulls (∼USD $100 Lifetime Net Merit index, NM$) ( Mueller et al., 2019 ).

Neither $JapOx nor NM$ incorporate polled directly into the selection index, although new horned calves were penalized with a 2% dehorning mortality rate in the current simulation study. Cole (2015) found that the inclusion of an economic weight on polled based on the cost of dehorning (including reduced calf health and increased calf mortality) had no effect on allele frequency in the United States Holstein population. Rather, the primary determinant of success with breeding for polled was the initial allele frequency in the population being studied. Adding a polled status weighting to $JapOx in the current study would not have altered the number of polled bulls available in scenarios B, C, and D and thus would not have altered the rate of HORNED allele frequency change in those scenarios. However, it would have increased the$JapOx value of polled sires which would have accelerated the decrease of the HORNED allele frequency in the baseline scenario A if the weighting exceeded the genetic merit differential between horned and homozygous polled bulls.

It should be noted that the average genetic merit values used in this study may not be completely representative of the Australian Brahman population as producers are unlikely to have sent in samples from horned animals for polled DNA testing. The data did suggest a slight over representation of heterozygotes, although the observed frequencies were not significantly different from those expected under Hardy–Weinberg equilibrium. However, if homozygous polled animals have a different genetic merit value profile in reality than were used in this study, then the results would alter accordingly.

### Mating Schemes

Several mating schemes were modeled in the current study to represent a business-as-usual scenario and potential situations that might arise if some, or drastic (i. e., prohibited production of horned calves) market pressure is placed on the Australian beef industry to eliminate dehorning. In our simulation results, both the seedstock and commercial population results for each scenario followed similar patterns. Although each scenario had a greater impact on the seedstock population due to its smaller size, an important measure of a seedstock population is its gene flow into the larger commercial population. Therefore, the following discussion will focus primarily on the commercial population results.

#### Baseline (A)

Baseline scenario A placed no selection pressure on polled genetics, which represents industry practices where breeders are solely focused on improving genetic merit ($JapOx). This scenario resulted in little change in HORNED allele frequency ( Figures 5A, B ) or genotypes of sires used for breeding ( Figure 6 ). As mentioned, there is a larger proportion of polled Australian Brahman sires compared to United States dairy sires. Irrespective, the baseline scenario demonstrated that, as with the United States dairy industry ( Cole, 2015 ; Mueller et al., 2019 ), strong selection pressure will need to be placed on polled genetics in order to meaningfully increase the number of polled Australian Brahman cattle. The baseline scenario A commercial population reached a 20-year$JapOx value of $160 ($8/year; Figure 5F ). The baseline scenario commercial population gain of $8 per year is higher than the current rate of gain for the Australian Brahman population ( Australian Brahman Breeder’s Association [ABBA], 2018 ). One reason for this high rate of gain is that this simulation assumed that TBV for$JapOx were known (i. e., breeding value accuracy = 1). Additionally, no regard was given to practical considerations and other selection criteria such as temperament, phenotype, and structure, which would reduce the rate of gain, and also influence selection for poll/edited animals and therefore decrease the rate of actual population change in polled. Conversely, no selection was placed on females in this simulation. Voluntary culling of horned cows would be another way to reduce the HORNED allele frequency, although it may limit the number of cows actually available to breed in some scenarios and make little difference in others. For example, in scenario D, due to population size, there is no involuntary culling of cows until year 15, and by that time there are no horned animals remaining to cull.

Baseline scenario A demonstrates the trade-offs associated with the business-as-usual situation. While scenario A resulted in one of the fastest rates of genetic gain, this scenario also used the most horned bulls for breeding throughout all 20 simulated years ( Figures 7A, B ) and consequently produced the most horned calves. Due to the northern Australian environment and extensive management practices there is a ∼2% mortality rate associated with dehorning ( Bunter et al., 2013 ). This dehorning mortality rate was included in our model and resulted in scenario A having the smallest number of total animals sold for beef in year 20, which is a major opportunity cost associated with continuing to produce horned calves ( Figure 8B ).

#### Polled Preference (B and C)

The polled preference scenarios were chosen to represent an intermediate situation where some producers prioritize using polled sires (scenario B) or choose to only use homozygous polled sires (scenario C) while other producers continue to select sires based solely on $JapOx. This situation may arise if AHA were to implement the public suggestion of “ mandating pain relief irrespective of the age of the animal” for surgical procedures, such as dehorning ( Animal Health Australia [AHA], 2014b ). Both scenario B and C resulted in a similar decrease of commercial HORNED allele frequency to ∼30% by year 20. Interestingly, the Australian Brahman homozygous polled-only preference scenario C results contrast with the same scenario modeled for the United States dairy population. In the United States dairy simulations, scenario C (i. e., preferentially using homozygous polled sires) did not significantly reduce the HORNED allele frequency because there were not enough homozygous polled Holstein sires available (only 1% of total sires) for scenario C to be successful. In contrast, there is a larger proportion of Australian Brahman homozygous polled sires available (6% of total sires), so C was the optimal conventional breeding scenario for the Brahman population in terms of substantially decreasing HORNED allele frequency while sustaining the rate of genetic gain. This simulation finding is in agreement with Lyons and Randhawa (2020). When polled and horned cohorts’ BREEDPLAN EBVs were compared, they concluded, “ An increased prevalence of the polled condition within herd or industry-wide would not be expected to have a measurable negative influence on production, carcass, fertility and behavior traits” ( Lyons and Randhawa, 2020 ). As demonstrated in polled preference scenarios, B and C, Australian Brahman producers currently have genetic options available to proactively address this animal welfare concern before it becomes a legal requirement. In our simulation there was a significant decrease ($1. 30/year) in the rate of genetic gain in conventional breeding scenario B and C compared to the baseline scenario A. However, after 10 years of both B and C there was a benefit in that these scenarios had less dehorning-related calf loss resulting in over 200 more steers sold for beef annually as compared to scenario A.

Additionally, producing high-genetic-merit polled sires through gene editing could mitigate the economic trade-off of the slower rate of genetic gain associated with increasing POLLED allele frequency through conventional breeding methods. Consistent with United States dairy findings ( Mueller et al., 2019 ), the scenarios that placed strong selection pressure on polled (scenarios B and C) combined with the use of gene editing, were the optimum solution for decreasing the number of horned calves, while still maintaining the rate of genetic gain.

In the United States dairy population, the addition of gene editing only the top 1% of genetic merit bull calves per year to the polled preference schemes was sufficient to maintain the same or better rate of genetic gain compared to baseline A, while significantly decreasing horned to less than 10% ( Mueller et al., 2019 ). Due to the widespread use of AI, a single dairy sire can have an immense impact on the whole population, thus only a small number of elite dairy sires need to be gene edited to see population level results.

In contrast, due to natural mating limits, individual Australian Brahman gene edited bulls did not have an extensive impact on the whole population. While the addition of gene editing the top 1% of Australian Brahman seedstock bull calves per year did significantly increase the rate of genetic gain ($0. 80/year) compared to the conventional breeding scenarios, scenarios B_1% and C_1% still resulted in significantly slower ($0. 50/year) rates of genetic gain compared to baseline A. However, the addition of gene editing the top 10% of Australian Brahman seedstock bull calves per year did result in similar rates of genetic gain to baseline A. Therefore, the overall optimal Australian Brahman scenario based on maintaining the rate of gene gain and rapidly decreasing HORNED allele frequency was C_10%.

#### Obligatory Polled (D)

Obligatory polled scheme D modeled a case where market expectations or government policies force the Australian beef industry to completely ban dehorning, as previously advocated for by animal rights groups ( Animal Health Australia [AHA], 2014b ). This situation would require either the exclusive use of homozygous polled sires to ensure no horned progeny were produced or leaving all horns intact. Although, it is worth noting that a requirement for live export is cattle being polled or dehorned such that each horn stump is less than 12 cm in length and fully healed ( LiveCorp-Australia, 2001 ).

We chose to model the use of commercial homozygous polled sires to supplement the supply of seedstock bulls when necessary. This model could be achieved by separating and leaving intact (i. e., not castrating) any commercial bulls without horns during the first calf gathering and processing. These phenotypically hornless commercial bulls could then be genomically tested to determine their polled genotype. The confirmed homozygous polled commercial bulls could be retained or sold for breeding while the heterozygous polled commercial bulls could be castrated and sold for beef as usual.

The obligatory polled conventional breeding scenario D, resulted in the lowest 20-year HORNED allele frequency of less than 10%, but scenario D also had the slowest rate of genetic gain of $5. 50/year, compared to baseline scenario A of$8. 00/year. Although there was a significant difference between the Australian Brahman obligatory polled conventional breeding scenario D and baseline A rates of genetic gain (∼$2. 50/year), the difference was much less drastic than between the similar United States dairy scenarios (∼$34. 00/year) ( Mueller et al., 2019 ). Additionally, unlike United States dairy simulations, the obligatory polled scheme D did not increase inbreeding above acceptable levels. Instead the primary issue for Australian Brahman’s in this scenario was not having a sufficient number of homozygous polled bulls to breed all commercial females via natural service, even when supplementing the seedstock supply of bulls with bulls sourced from the commercial sector.

In this simulation, once all the seedstock and commercial homozygous polled bulls reached their maximum 35 matings per year, the remaining females were left open. This model was chosen to demonstrate the current deficit of homozygous polled Australian Brahman bulls that would be needed to make this obligatory polled scheme D feasible. A more realistic alternative would be to use homozygous polled bulls from other less tropically adapted breeds. However, this alternative could result in increased animal welfare issues and higher levels of mortality due to ill-adapted sires and their resultant progeny ( Bunter et al., 2013 ).

### Expert paper writers are just a few clicks away

Place an order in 3 easy steps. Takes less than 5 mins.

error: Content is protected !!
MyNursingHomeworkHelp
Need Help With Your Assignment? Lets Talk