Genetic and phenotypic correlations between Aleutian disease tests with body weight, growth, and feed efficiency traits in mink

Abstract

The ineffectiveness of vaccination, medicine, and culling strategy leads mink farmers to control Aleutian disease (AD) by selecting AD-resilient mink based on AD tests. However, the genetic background of AD tests and their correlations with economically important or AD-resilient traits are limited. This study estimated the genetic and phenotypic correlations between four AD tests and seven body weight (BW) traits, six growth parameters from the Richards growth model, and eight feed-related traits. Univariate models were used to test the significance (P < 0.05) of fixed effects (sex, color type, AD test year, birth year, and row-by-year), random effects (additive genetic, maternal genetic, and permanent environmental), and a covariate of age using ASReml 4.1. Likewise, pairwise bivariate analyses were conducted to estimate the phenotypic and genetic correlations among the studied traits. Both antigen- and virus capsid protein-based enzyme-linked immunosorbent assay tests (ELISA-G and ELISA-P) showed significant (P < 0.05) moderate positive genetic correlations (±SE) with maturation rate (from 0.36 ± 0.18 to 0.38 ± 0.19). ELISA-G showed a significant negative genetic correlation (±SE) with average daily gain (ADG, −0.37 ± 0.16). ELISA-P showed a significant positive moderate genetic correlation (±SE) with off-feed days (DOF, 0.42 ± 0.17). These findings indicated that selection for low ELISA scores would reduce the maturation rate, increase ADG (by ELISA-G), and minimize DOF (by ELISA-P). The iodine agglutination test (IAT) showed significant genetic correlations with DOF (0.73 ± 0.16), BW at 16 weeks of age (BW16, 0.45 ± 0.23), and BW at harvest (HW, −0.47 ± 0.20), indicating that selection for lower IAT scores would lead to lower DOF and BW16, and higher HW. These estimated genetic correlations suggested that the selection of AD tests would not cause adverse effects on the growth, feed efficiency, and feed intake of mink. The estimates from this study might strengthen the previous finding that ELISA-G could be applied as a reliable and practical indicator trait in the genetic selection of AD-resilient mink in AD-positive farms.

The estimates from this study indicated the selection of Aleutian disease tests would not cause adverse effects on growth, feed intake, and feed efficiency in mink. Meanwhile, the estimates further illustrate the antigen-based enzyme-linked immunosorbent assay test is the most reliable and practical indicator among all Aleutian disease tests in selecting Aleutian disease-resilient mink.

Introduction

As one of the most severe diseases in mink farming, Aleutian disease (AD) brings tremendous financial losses to the mink industry and has made it difficult for mink farmers to maintain their farming. Aleutian disease is caused by the Aleutian mink disease virus (AMDV). AMDV is a non-enveloped single-stranded DNA virus in parvovirus family (genus Amdoparvovirus, species Carnivore amdoparvovirus 1). AMDV infection causes significant pathology, including hypergammaglobulinemia, glomerulonephritis, plasmacytosis, and arteritis to the infected mink (Porter et al., 1969, 1973; Cho and Ingram, 1973). AMDV infection also causes adverse effects on female reproductive performance for example small litter size, fetal death, and abortion (Henson et al., 1962; Reichert and Kostro, 2014). Additionally, small body size due to such as chronic progressive weight loss (Porter et al., 1982), low feed intake (inappetence) (Eklund et al., 1968; Jensen et al., 2016), and poor pelt quality characterized by hair depigmentation (Farid and Ferns, 2011) are also adverse outcomes. Together, these clinical features on AMDV infection in farmed mink directly reduce the financial income of mink farmers. Therefore, an effective, practical, and reliable approach is urgently needed by the mink farmers to minimize the adverse effects of AD. Unfortunately, vaccination and recovery from natural infection have not been identified as viable strategies of protection due to antibody disease enhancement. Specifically, antibodies elicited toward the Aleutian mink disease virus are not capable of neutralizing the virus and instead the infectious virus forms complexes with pre-existing antibody complexes leading to lesions in the glomerulus and arteries of the mink host (Porter et al., 1969, 1973; Cho and Ingram, 1973; Stolze and Kaaden, 1987).

Several methods have been attempted to control AD for farmed mink, however, at this time control measures have not been successful. No effective vaccine or medicine has been developed with the ability to protect against infection without inducing vaccine-enhanced disease (Liu et al., 2017; Farid et al., 2018). Culling mink after testing positive for AD has been the primary method to control AD, however, culling does not successfully prevent future outbreaks (Gunnarsson, 2001; Christensen et al., 2011; Themudo et al., 2011; Farid et al., 2012). Feeding mink, which were challenged with AMDV, with dietary kelp (Ascophylum nodosum) supplementation showed some benefits to the survival rate of adult mink and litter size, but the effects were not significant compared to the control group (Farid et al., 2020). In recent years, the selection of AD-resilient individuals based on AD tests is suggested as a potential solution to cope with the adverse effects caused by AD effectively (Farid et al., 2018; Hu et al., 2020).

Several AD tests have been applied in the phenotypic selection of AD-resilient mink in a number of mink-producing areas, but the feasibility of employing AD tests in the genetic selection of AD-resilient mink has not been evaluated. Several mink farms in the province of Nova Scotia in Canada selected \AD-resilient mink based on the productive performance and iodine agglutination test (IAT, measuring gamma globulin level) results (Farid and Ferns, 2017). Some AD positive mink farms in North America and Europe (Knuuttila et al., 2009; Farid and Rupasinghe, 2016; Farid et al., 2018) have selected AD-resilient mink based on enzyme-linked immunosorbent assay tests (ELISA, measuring anti-AMDV antibody level). Although some mink farmers use AD tests in the phenotypic selection of AD-resilient mink, the feasibility of this approach has not been explored due to the lack of comprehensive genetic background knowledge of AD tests.

An effective genetic selection program requires a comprehensive study of genetic and phenotypic parameters for traits of interest (Toghiani, 2012). Studies of the genetic parameter of AD tests and their correlations with AD-resilient or other economically important traits are rare. However, Hu et al (2021) estimated heritabilities (±SE) for AMDV antigen-based ELISA (ELISA-G), AMDV capsid protein-based ELISA (ELISA-P), counterimmunoelectrophoresis (CIEP), and IAT of 0.39 ± 0.06, 0.61 ± 0.07, 0.11 ± 0.07, and 0.26 ± 0.05, respectively. The genetic and phenotypic correlations between AD tests with pelt quality, reproductive performance, packed-cell volume, and harvest length were also estimated by (Hu et al., 2021), and ELISA-G was suggested as a good indicator trait for selecting AD-resilient mink. A comprehensive view of the genetic and phenotypic correlations between target traits and other important traits is important for genetic selection. Here in our current study, we further investigated the genetic and phenotypic correlations between AD tests and other economically important or AD-resilient traits to provide a more comprehensive view of the genetic and phenotypic correlations between AD tests and other economically important traits in mink. The overall objective of this study was to estimate the genetic and phenotypic correlations between AD tests with body weights at different ages, growth parameters obtained from the Richards growth model (Do and Miar, 2019), different measurements of feed efficiency, and feed-intake-related traits.

Materials and Methods

This study was approved by the Dalhousie University Animal Care and Use Committee (certification#: 2018-009 and 2019-012). All the mink were raised based on the Code of Practice for the Care and Handling of Farmed Mink guidelines from the Canada Mink Breeders Association (NFACC, 2013).

Animals and management

Animals in this study were raised under standard farming conditions at the Canadian Centre for Fur Animal Research (CCFAR) at Dalhousie University, Faculty of Agriculture (Truro, Nova Scotia, Canada) from 2013 to 2021. All mink had ad libitum access to food and water. The annual mink production cycle in CCFAR contains four periods: 1) conditioning and breeding season (December–March), where mink farmers adjust the feed to provide the mink with good conditions for breeding in March; 2) whelping and weaning season (April–June), where females will give birth at the end of April or early May, and the kits will be weaned at the end of June; 3) growth and furring season (July–October), animals will be fed ad-lib to fulfill their growth potential; and 4) grading and harvesting (November–December), where farmers determine which mink will be pelted or kept as breeding stock for the future season. An AD outbreak of CCFAR was identified in 2013. The source of this outbreak was not determined. AMDV-contaminated feed as well as contact with wild animals carrying AMDV have been considered as the most likely causes. Meanwhile, no persistent breeding program was employed in CCFAR during the study years (2013–2021). The total of 2,488 mink (males = 832 and females = 1,656) used in this study were the progeny of 444 sires and 852 dams. Pedigree information of 17 generations comprising 24,864 individuals was used.

Aleutian disease tests

All AD tests were conducted using established protocols described by Hu et al. (2021). Briefly, blood samples of the studied mink (n = 2,352) were collected using the toenail clipping approach at two different periods: 1) in mid-November of 2013, 2014, 2018, and 2019 before selecting breeders; and 2) in mid-February of 2013, 2014, 2017, 2018, 2019, 2020, and 2021 before mating. Both ELISA-G and ELISA-P systems were employed to quantify the anti-AMDV antibodies in the serum. The ELISA-G test was conducted at Middleton Veterinary Services (Nova Scotia, Canada), and the test results included eight categories from 0 (low) to 7 (high), with 1 point increments. The ELISA-P tests were conducted at Nederlandse Federatie van Edelpelsdierenhouders (Wijchen, Netherlands), and the test results included nine categories from 0 (low) to 8 (high), with 1 point increments. The CIEP tests were conducted at the Animal Health Laboratory at the University of Guelph (Guelph, Canada) to detect the existence of anti-AMDV antibodies in the blood samples, and the results were recorded as 0 (negative) and 1 (positive). The IAT tests were conducted at CCFAR to measure the serum gamma globulin level in the serum, and the results were scored into four categories from 0 (low) to 4 (high).

Body weight measurements and growth parameters

A total of 1,088 mink born in 2018 and 2019 were randomly selected for collecting body weight (BW) data. All BW data were collected using established protocols described by Do et al. (2021). Briefly, selected mink were weighed at birth, which happens between the end of April and early May, and at weaning around the end of June (about seven weeks of age). After weaning, each selected mink was raised in a single cage and weighed every three weeks from 13 to 28 weeks after birth. Thus, the BW traits included BW at week 13 (BW13), week 16 (BW16), week 19 (BW19), week 22 (BW22), week 25 (BW25), and week 28 (BW28). The BW of mink at harvest (HW) was measured at two different times. The HW of mink, which were not selected as breeders, were measured at harvest days in December 2018 and 2019. Additionally, HW of sires that completed their breeding tasks and dams that mated but failed to become pregnant were measured at harvest days in February 2019 and 2020.

Growth parameters were derived from the Richards growth model (Richards and Kavanagh, 1945). This model was suggested previously to be the most suitable model to describe the growth of mink (Liu et al., 2011; Do and Miar, 2019). The minpack.lm packages (Elzhov et al., 2016) in R software (R Development Core Team, 2011) was used to fit individual BW records into the following Richards growth model, as described by Do et al. (2021):

$${\displaystyle \begin{array}{l}{\displaystyle BWt=\frac{\alpha }{{(1-\beta \times e^{-kt})}^{1/m}},}\end{array}}$$

where BWt is the BW (kg) at age t (weeks), α is the asymptotic weight (kg), β is the parameter characterizing the first part of growth before the inflection point, k is the maturation rate, and m is the inflection parameter.

The parameters, α, m, and k, derived from Richards model were then used to calculate the age at the inflection point (AIP) and the weight at the inflection point (WIP) as follows:

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{A}\mathrm{I}\mathrm{P}=\frac{\alpha }{{(m+1)}^{1/m}},}\end{array}}$$

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{W}\mathrm{I}\mathrm{P}=\frac{-In\left(m/\beta \right)}{k}}\end{array}}$$

Feed intake measurement and feed efficiency

The feed intake data were collected from August 1st to November 14th in 2018 and 2019 at CCFAR using the established protocols described by Davoudi et al. (2022). Briefly, 1,088 mink had both feed intake and growth traits data. Mink were housed individually in single cages, and feed was distributed to each pen every day. The amounts of feed allocated to mink were regulated based on the leftover records to avoid unnecessary feed waste and meet the mink’s appetite. The difference between the amount of feed leftover and feed provided was recorded as daily feed intake (DFI). The average of the DFI records during the feeding trial was used as the average daily feed intake (ADFI).

Feed efficiency traits were calculated using both feed intake and growth data. The BW13 was treated as the initial BW, and the BW on the last day of the feeding trial was treated as the final BW. The following equation was used to calculate the average daily gain (ADG):

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{A}\mathrm{D}\mathrm{G}=\frac{\mathrm{F}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}\mathrm{B}\mathrm{W}-\mathrm{I}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l}\mathrm{B}\mathrm{W}}{\mathrm{N}\mathrm{u}\mathrm{m}\mathrm{b}\mathrm{e}\mathrm{r}\mathrm{o}\mathrm{f}\mathrm{d}\mathrm{a}\mathrm{y}\mathrm{s}\mathrm{o}\mathrm{n}\mathrm{t}\mathrm{h}\mathrm{e}\mathrm{t}\mathrm{e}\mathrm{s}\mathrm{t}},}\end{array}}$$

For calculating the feed conversion ratio (FCR), the following equation was used:

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{F}\mathrm{C}\mathrm{R}=\frac{\mathrm{A}\mathrm{D}\mathrm{F}\mathrm{I}}{\mathrm{A}\mathrm{D}\mathrm{G}},}\end{array}}$$

The mid-test metabolic BW (BW^0.75) and Kleiber ratio (KR) were calculated using the following equations, respectively:

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{B}{\mathrm{W}}^{0.75}={\left(\frac{\mathrm{I}\mathrm{n}\mathrm{i}\mathrm{t}\mathrm{i}\mathrm{a}\mathrm{l}\mathrm{B}\mathrm{W}+\mathrm{F}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}\mathrm{B}\mathrm{W}}{2}\right)}^{0.75},}\end{array}}$$

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{K}\mathrm{R}=\frac{\mathrm{A}\mathrm{D}\mathrm{G}}{\mathrm{B}{\mathrm{W}}^{0.75}}}\end{array}}$$

The linear regression model used to estimate the residual feed intake (RFI) was:

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{A}\mathrm{D}\mathrm{F}\mathrm{I}={\beta }_0+{\beta }_1\mathrm{A}\mathrm{D}\mathrm{G}+{\beta }_2\mathrm{B}{\mathrm{W}}^{0.75}+\epsilon ,}\end{array}}$$

where β₀ is the equation intercept, β₁ is the partial regression coefficient of ADG, β₂ is the partial regression coefficient of BW^0.75, and ε is RFI.

Residual gain (RG) was estimated using the following linear regression model:

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{A}\mathrm{D}\mathrm{G}={\beta }_0+{\beta }_1\mathrm{A}\mathrm{D}\mathrm{F}\mathrm{I}+{\beta }_2\mathrm{B}{\mathrm{W}}^{0.75}+\epsilon ,}\end{array}}$$

where β₀ is the equation intercept, β₁ is the partial regression coefficient of ADFI, β₂ is the partial regression coefficient of BW^0.75, and ε is RG.

The following equation (Berry and Crowley, 2012) was used to derive the residual intake and gain (RIG):

$${\displaystyle \begin{array}{l}{\displaystyle \mathrm{R}\mathrm{I}\mathrm{G}=(-1\times \left(\frac{\mathrm{R}\mathrm{F}{\mathrm{I}}_i}{\mathrm{R}\mathrm{F}{\mathrm{I}}_{sd}}\right))+\left(\frac{\mathrm{R}{\mathrm{G}}_i}{\mathrm{R}{\mathrm{G}}_{sd}}\right),}\end{array}}$$

where RFI_i, is the RFI for individual i, RFI_sd is the standard deviation of RFI for all animals, RG_i is the residual gain for animal i, and RG_sd is the standard deviation of RG for all animals.

Feed-intake-relative traits

Two feed intake traits, the day-to-day variation in feed intake (VarF) and the proportion of off-feed days based on feed intake (DOF), were derived from the DFI data available for each mink using the method described by Putz et al. (2019). The VarF was measured based on the root mean square error of within-individual regression of DFI on day using ordinary least squares linear regression. The off-feed days were identified using a 5% quantile regression of DFI on age across all mink. The negative residuals below the regression line were treated as off-feed days for each mink. The DOF was calculated as the proportion of off-feed days to the total recorded days of DFI.

Statistical analyses

A univariate animal model was used to estimate the variance components of random additive genetic, permanent environmental, and maternal genetic effects for individual traits using ASReml 4.1 software (Gilmour et al., 2018). The model was as follows:

$${\displaystyle \begin{array}{l}{\displaystyle y=Xb+Za+Wpe+Gm+e,}\end{array}}$$

where y is the vector of phenotypes; X is the incidence matrix relating phenotypes to fixed effects;b is the vector of fixed effects; Z is the incidence matrix relating phenotypes to random additive genetic effects;a is the vector of random additive genetic effects, with $a\sim N(0,A{\sigma }_a^2)$, where A is the numerator relationship matrix, and ${\sigma }_a^2$ is the additive genetic variance; W is the incidence matrix relating phenotypes to random permanent environmental effects;pe is the vector of random permanent environmental effects, with $pe\sim N(0,I{\sigma }_{pe}^2)$, where I is an identity matrix, and ${\sigma }_{pe}^2$ is the permanent environmental variance; G is the incidence matrix relating phenotypes to random maternal genetic effects; m is the vector of random maternal genetic effects, with $m\sim N(0,A{\sigma }_m^2)$, where ${\sigma }_m^2$ is the maternal genetic variance; and e is the vector of residual effects, with $e\sim N(0,I{\sigma }_e^2)$, where ${\sigma }_e^2$ is the residual variance.

The significance of fixed effects and covariates was tested using Wald statistics in the REML procedure of ASReml 4.1 (Gilmour et al., 2018), and only significant (P < 0.05) effects were kept in the mixed model analyses for each trait (Table 1). The fixed effects of sex and color type were tested for all studied traits. The fixed effect of AD test year (2013 to 2021), which described the year that the animals were tested for AD, was tested for AD test traits. The fixed effect of row-by-year, which described the location of cages (six rows) where mink were raised in 2018 and 2019, was tested for feed efficiency and feed intake traits. The fixed effects of dam age (1–3 years old), which described the age of the animals’ dams when they were born, and birth year (2018 and 2019), which described the year that the animals were born, were tested for growth traits. Blood collection age (in days) obtained for AD tests was only tested for AD traits. The age at harvest (in days) was tested for harvest weight only. The age at the end of feeding trial (in days) was tested for both feed efficiency and feed intake traits.

Table 1.

Significance of fixed and random effects included in the models for analysis of Aleutian disease tests, growth, feed efficiency, and feed intake traits in mink

Traits1	Factors						Covariates			Random effects
	Sex	Color type	AD test year	Row-by-year	Dam age	Birth year	Blood test age	Harvest age	Feeding trial end age	Maternal genetic	Permanent environmental
ELISA-G	*	NS	*	-	-	-	NS	-	-	*	*
ELISA-P	NS	NS	NS	-	-	-	NS	-	-	*	NS
CIEP	*	NS	NS	-	-	-	*	-	-	*	NS
IAT	NS	NS	*	-	-	-	NS	-	-	NS	*
BW13	*	NS	-	-	NS	*	-	-	-	*	-
BW16	*	NS	-	-	NS	*	-	-	-	*	-
BW19	*	NS	-	-	NS	NS	-	-	-	*	-
BW22	*	NS	-	-	NS	NS	-	-	-	*	-
BW25	*	NS	-	-	NS	*	-	-	-	*	-
BW28	*	NS	-	-	NS	*	-	-	-	*	-
HW	*	*	-	-	NS	*	-	NS	-	*	-
ADG	*	NS	-	NS	-	-	-	-	*	*	-
α	*	NS	-	-	NS	NS	-	-	-	*	-
k	NS	NS	-	-	NS	*	-	-	-	*	-
m	NS	NS	-	-	NS	*	-	-	-	*	-
AIP	*	NS	-	-	NS	*	-	-	-	*	-
WIP	*	NS	-	-	NS	NS	-	-	-	*	-
DFI	*	NS	-	*	-	-	-	-	NS	*	-
FCR	*	NS	-	*	-	-	-	-	*	*	-
RFI	*	NS	-	*	-	-	-	-	NS	*	-
RG	NS	NS	-	NS	-	-	-	-	*	*	-
RIG	*	NS	-	*	-	-	-	-	*	*	-
KR	*	NS	-	*	-	-	-	-	*	*	-
DOF	NS	NS	-	*	-	-	-	-	*	NS	-
VarF	*	NS	-	*	-	-	-	-	*	NS	-

¹ELISA-G, AMDV-G based enzyme-linked immunosorbent assay test; ELISA-P, VP2 based enzyme-linked immunosorbent assay test; CIEP, counterimmunoelectrophoresis test; IAT, iodine agglutination test; BW (13–28), body weight at corresponding measurement week (13–28); HW, body weight at harvest; ADG, average daily gain; α, asymptotic weight in kg; k, second part of growth curve, in which growth rate decreases until the animal reaches the asymptotic or mature weight (α); m, shape parameter determining the position of curve inflection point; AIP, age at the inflection point; WIP, weight at the inflection point; DFI, daily feed intake; FCR, feed conversion ratio; RFI, residual feed intake; RG, residual gain; RIG, residual intake and gain; KR, Kleiber ratio; DOF, the proportion of off-feed days; VarF, variation in daily feed intake.

^*Significant (P < 0.05).

NS = not significant (P > 0.05).

-not tested.

The following likelihood ratio tests, which compares the full model and the reduced model, was used to test the significance of the random maternal genetic and permanent environmental effects:

$${\displaystyle \begin{array}{l}{\displaystyle -2(\mathrm{l}\mathrm{o}\mathrm{g}{L}_{\mathrm{r}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{e}\mathrm{d}\mathrm{m}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l}}-\mathrm{l}\mathrm{o}\mathrm{g}{L}_{\mathrm{f}\mathrm{u}\mathrm{l}\mathrm{l}\mathrm{m}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l}})}\\ \mathrm{x}2\mathrm{d}\mathrm{f}(\mathrm{f}\mathrm{u}\mathrm{l}\mathrm{l}\mathrm{m}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l})-\mathrm{d}\mathrm{f}(\mathrm{r}\mathrm{e}\mathrm{d}\mathrm{u}\mathrm{c}\mathrm{e}\mathrm{d}\mathrm{m}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l})\end{array}}$$

where log L and df are log-likelihood and degrees of freedom. The random maternal genetic effect was tested for all studied traits. The random permanent environmental effect was only tested for AD tests because only AD tests had repeated records in different years.

To estimate the genetic and phenotypic correlations between traits, the bivariate models were conducted using ASReml 4.1 software (Gilmour et al., 2018). Relevant significant fixed and random effects were included in bivariate analyses for each trait (Table 1). The following bivariate model was used to analyze the traits:

$${\displaystyle \begin{array}{l}{\displaystyle \left[\begin{array}{l}{y}_1\\ y_2\end{array}\right]=\left[\begin{array}{ll}{X}_1& {\displaystyle 0}\\ 0& {\displaystyle X_2}\end{array}\right]\left[\begin{array}{l}{b}_1\\ b_2\end{array}\right]+\left[\begin{array}{ll}{Z}_{a1}& {\displaystyle 0}\\ 0& {\displaystyle Z_{a2}}\end{array}\right]\left[\begin{array}{l}{a}_1\\ a_2\end{array}\right]}\\ +\left[\begin{array}{ll}{Z}_{pe1}& {\displaystyle 0}\\ 0& {\displaystyle Z_{pe2}}\end{array}\right]\left[\begin{array}{l}p{e}_1\\ p{e}_2\end{array}\right]+\left[\begin{array}{ll}{Z}_{m1}& {\displaystyle 0}\\ 0& {\displaystyle Z_{m2}}\end{array}\right]\left[\begin{array}{l}{m}_1\\ m_2\end{array}\right]+\left[\begin{array}{l}{e}_1\\ e_2\end{array}\right],\end{array}}$$

where subscript numbers 1 and 2 refer to trait 1 and 2; y is the vector of phenotypes; X is the incidence matrix relating phenotypes to fixed effects; b is the vector of fixed effects; Z_a is the incidence matrix relating phenotypes to random additive genetic effects; a is the vector of random additive genetic effects; Z_pe is the incidence matrix relating phenotypes to random permanent environmental effects; pe is the vector of random permanent environmental effects; Z_m is the incidence matrix relating phenotypes to random maternal genetic effects; m is the vector of random maternal genetic effects; and e is the vector of residual effects. The assumptions for the bivariate random effects were:

$${\displaystyle \begin{array}{l}{\displaystyle \left[\begin{array}{l}{a}_1\\ a_2\end{array}\right]\sim MVN(0,A\otimes \left[\begin{array}{cc}{\sigma }_{a1}^2& {\displaystyle {\sigma }_{a1a2}}\\ {\sigma }_{a1a2}& {\displaystyle {\sigma }_{a2}^2}\end{array}\right]),}\end{array}}$$

$${\displaystyle \begin{array}{l}{\displaystyle \left[\begin{array}{l}p{e}_1\\ p{e}_2\end{array}\right]\sim MVN(0,I\otimes \left[\begin{array}{cc}{\sigma }_{pe1}^2& {\displaystyle {\sigma }_{pe1pe2}}\\ {\sigma }_{pe1pe2}& {\displaystyle {\sigma }_{pe2}^2}\end{array}\right]),}\end{array}}$$

$${\displaystyle \begin{array}{l}{\displaystyle \left[\begin{array}{l}{m}_1\\ m_2\end{array}\right]\sim MVN(0,A\otimes \left[\begin{array}{cc}{\sigma }_{m1}^2& {\displaystyle {\sigma }_{m1m2}}\\ {\sigma }_{m1m2}& {\displaystyle {\sigma }_{m2}^2}\end{array}\right]),}\end{array}}$$

and

$${\displaystyle \begin{array}{l}{\displaystyle \left[\begin{array}{l}{e}_1\\ e_2\end{array}\right]\sim MVN(0,I\otimes \left[\begin{array}{cc}{\sigma }_{e1}^2& {\displaystyle {\sigma }_{e1e2}}\\ {\sigma }_{e1e2}& {\displaystyle {\sigma }_{e2}^2}\end{array}\right]),}\end{array}}$$

where subscript numbers 1 and 2 refer to traits 1 and 2; A is the numerator relationship matrix; I is an identity matrix; ${\sigma }_a^2$ is the variance of random additive genetic effects; ${\sigma }_{a1a2}$ is the covariance between random additive genetics effects for traits 1 and 2; ${\sigma }_{pe}^2$ is the variance of random permanent environmental effects; ${\sigma }_{pe1pe2}$ is the covariance between random permanent environmental effects for traits 1 and 2; ${\sigma }_m^2$ is the variance of random maternal genetic effects; ${\sigma }_{m1m2}$ is the covariance between random maternal genetic effects for traits 1 and 2; ${\sigma }_e^2$ is the variance of random residual effects; and ${\sigma }_{e1e2}$ is the covariance between residual effects for traits 1 and 2.

Random additive genetic effects were included in the final model for all traits. Random permanent environmental effect was only significant (P < 0.05) for ELISA-G and IAT (Table 1). Random maternal genetic effect was included in the final models for all studied traits except for IAT, DOF, and VarF, because the random maternal genetic effects were not significant (P >0.05) for these traits (Table 1).

Phenotypic variances were calculated as ${\sigma }_p^2$ = ${\sigma }_a^2$ + ${\sigma }_e^2$ for DOF and VarF, as ${\sigma }_p^2$ = ${\sigma }_a^2$ + ${\sigma }_m^2$ + ${\sigma }_e^2$ for ELISA-P, CIEP, all growth traits, and all feed efficiency traits, as ${\sigma }_p^2$ = ${\sigma }_a^2$ + ${\sigma }_{pe}^2$ + ${\sigma }_e^2$ for IAT, and as ${\sigma }_p^2$ = ${\sigma }_a^2$ + ${\sigma }_{pe}^2$ + ${\sigma }_m^2$ + ${\sigma }_e^2$ for ELISA-G. Heritability (h²) was calculated as follows:

$${\displaystyle \begin{array}{l}{\displaystyle h^2=\frac{{\sigma }_a^2}{{\sigma }_p^2},}\end{array}}$$

Repeatability (r²) was defined as follows:

$${\displaystyle \begin{array}{l}{\displaystyle r^2=\frac{{\sigma }_a^2+{\sigma }_{pe}^2}{{\sigma }_p^2}}\end{array}}$$

Phenotypic and genetic correlations among traits were calculated based on bivariate models’ (co)variance components.

In this study, the binary trait of CIEP was analyzed as a continuous trait, which was not theoretically optimal as the suitable methodology was the threshold model (Gianola, 1982). The generalized linear mixed models (GLMM) procedures with animal model in ASReml (Gilmour et al., 2018) were initially used to analyze the CIEP trait, but this method failed to estimate the variances caused by random additive genetic effects. The GLMM procedure with sire model was suggested as a more appropriate method to run threshold model on the user guide of ASReml (Southey and Gilmour, 2021). However, the sire model was not practical for mink breeding because the current breeding procedure in mink had more focused on dams than sires. In the meantime, several other studies also treated binary traits as continuous traits when they estimated genetic parameters for disease-related traits (Farid et al., 2018; Gunia et al., 2018). Thus, CIEP was treated and analyzed as a continuous trait in this study.

Results and Discussion

Descriptive statistics

The number of records, mean, SD, range, and coefficient of variation (CV) for each trait are presented in Table 2. Among all studied traits, ELISA-G (n = 2,896) and IAT (n = 2,236) had the most data records because CCFAR started recording these traits from 2013 while the other traits were measured after 2018. Compared with the previous study (Hu et al., 2021), we collected 537 more records of ELISA-G and 531 records of IAT in 2021. The lowest number of records was for HW (844) because this trait was only recorded after 2018. The CVs of ELISA-G (125%) and IAT (150.75%) were 12.33% larger and 12.41% lower than our previous study (Hu et al., 2021), respectively, which may be due to more records being used in this study. The CVs for BW traits at different ages increased with animal age and ranged from 23.39 to 31.28%. Except for growth parameter (m), the CVs of growth parameters were less than 42%, which ranged from 16.88% to 41.67%. The CVs ranged from 28.60% to 45.67% for feed efficiency traits. In addition, the distributions of all four AD test traits are presented in Figure 1.

Table 2.

Descriptive statistics for Aleutian disease tests, growth, feed efficiency, and feed intake traits in mink

Traits1	Number of records	Mean	SD	Range	CV (%)
ELISA-G	2,896	1.92	2.40	0–7	125.00
ELISA-P	1,152	2.14	2.16	0–8	100.93
CIEP	1,127	0.82	0.38	0–1	46.34
IAT	2,236	0.67	1.01	0–4	150.75
BW13 (kg)	1,059	1.24	0.29	0.64–1.95	23.39
BW16 (kg)	1,056	1.54	0.41	0.66–2.53	26.62
BW19 (kg)	1,051	1.77	0.53	0.75–3.03	29.94
BW22 (kg)	1,046	1.99	0.60	0.96–3.4	30.15
BW25 (kg)	1,032	2.07	0.63	0.93–3.75	30.43
BW28 (kg)	1,026	2.11	0.66	1.00–3.86	31.28
HW (kg)	844	2.09	0.64	0.94–3.94	30.62
ADG (g/day)	1,046	8.32	3.80	1.27–19.67	45.67
α (kg)	1,029	2.18	0.70	0.99–4.62	32.11
k	1,025	0.24	0.10	0.01–0.85	41.67
m	1,020	0.64	0.84	0.17 × 10−5 to 4.79	131.25
AIP (week)	1,018	10.90	1.84	6.40–17.72	16.88
WIP (kg)	1,024	0.97	0.35	0.13–2.20	36.08
DFI (g/day)	1,079	224.90	57.68	89.84–366.89	25.65
FCR	1,038	31.07	11.34	11.04–78.28	36.50
RFI (g/day)	1,044	−0.06	36.63	−86.57 to 169.42	-
RG (g/day)	1,042	0.01	1.37	−5.77 to 5.34	-
RIG (g/day)	1,043	−0.01	1.52	−5.61 to 5.08	-
KR	1,046	5.42	1.55	1.15–9.73	28.60
DOF	1,074	0.06	0.08	0.00–0.65	152.73
VarF	1,074	48.31	11.49	27.37–98.88	23.78

¹ELISA-G, AMDV-G based enzyme-linked immunosorbent assay test; ELISA-P, VP2 based enzyme-linked immunosorbent assay test; CIEP, counterimmunoelectrophoresis test; IAT, iodine agglutination test; BW (13–28), body weight at corresponding measurement week (13–28); HW, body weight at harvest; ADG, average daily gain; α, asymptotic weight in kg; k, second part of growth curve, in which growth rate decreases until the animal reaches the asymptotic or mature weight (α); m, shape parameter determining the position of curve inflection point; AIP, age at the inflection point; WIP, weight at the inflection point; DFI, daily feed intake; FCR, feed conversion ratio; RFI, residual feed intake; RG, residual gain; RIG, residual intake and gain; KR, Kleiber ratio; DOF, the proportion of off-feed days; VarF, variation in daily feed intake; SD, standard deviation; CV, coefficient of variation.

Figure 1.

The distribution of Aleutian disease tests. (A) VP2 based enzyme-linked immunosorbent assay test; (B) AMDV-G based enzyme-linked immunosorbent assay test; (C) counterimmunoelectrophoresis test; (D) iodine agglutination test.

Random maternal genetic and permanent environmental effects

Estimates of variance components and genetic parameters for all studied traits are presented in Table 3. Among AD tests, the random maternal genetic effect was significant (P < 0.05) for ELISA-G, ELISA-P, and CIEP (Table 1). In mink, the trans-placental transmission of AMDV was detected (Broll and Alexandersen, 1996), and thus, the significance of maternal genetic effect on AD-specific tests is expected. The random permanent environmental effect was significant (P < 0.05) for ELISA-G and IAT (Table 1). This was not the case for the latter in the previous study (Hu et al., 2021), which might be due to the larger sample size and more repeated IAT records on the same individuals in the current study.

Table 3.

Estimates of variance components and genetic parameters with their standard errors (SE) for Aleutian disease tests, growth, feed efficiency, and feed intake traits in mink

Traits1	Variance components				Genetic parameters
	${\sigma }_a^2$ ±SE	${\sigma }_m^2$ ±SE	${\sigma }_{pe}^2$ ±SE	${\sigma }_e^2$ ±SE	h 2 ± SE	$c_d^2$ ±SE	$c_{pe}^2$ ±SE	r 2 ± SE
ELISA-G	1.148 ± 0.245	0.585 ± 0.141	1.137 ± 0.168	1.501 ± 0.088	0.26 ± 0.05	0.13 ± 0.03	0.26 ± 0.40	0.52 ± 0.04
ELISA-P	2.950 ± 0.473	0.402 ± 0.204	NS	1.599 ± 0.075	0.59 ± 0.07	0.08 ± 0.04	NA	NA
CIEP	0.014 ± 0.010	0.020 ± 0.007	NS	0.114 ± 0.008	0.09 ± 0.07	0.14 ± 0.04	NA	NA
IAT	0.199 ± 0.044	NS	0.210 ± 0.051	0.597 ± 0.041	0.20 ± 0.04	NA	0.21 ± 0.51	0.41 ± 0.04
BW13	0.004 ± 0.002	0.006 ± 0.001	-	0.008 ± 0.001	0.23 ± 0.13	0.35 ± 0.07	NA	NA
BW16	0.008 ± 0.004	0.007 ± 0.002	-	0.017 ± 0.002	0.26 ± 0.12	0.23 ± 0.06	NA	NA
BW19	0.015 ± 0.006	0.008 ± 0.003	-	0.025 ± 0.003	0.31 ± 0.11	0.17 ± 0.05	NA	NA
BW22	0.024 ± 0.008	0.009 ± 0.003	-	0.036 ± 0.005	0.35 ± 0.11	0.13 ± 0.05	NA	NA
BW25	0.021 ± 0.009	0.010 ± 0.004	-	0.039 ± 0.005	0.30 ± 0.11	0.15 ± 0.05	NA	NA
BW28	0.019 ± 0.009	0.013 ± 0.004	-	0.044 ± 0.005	0.25 ± 0.11	0.17 ± 0.05	NA	NA
HW	0.013 ± 0.007	0.008 ± 0.004	-	0.062 ± 0.005	0.15 ± 0.09	0.10 ± 0.05	NA	NA
ADG	0.998 ± 0.434	0.610 ± 0.212	NS	2.645 ± 0.261	0.24 ± 0.10	0.14 ± 0.05	NA	NA
Α	0.016 ± 0.008	0.015 ± 0.005	-	0.066 ± 0.005	0.17 ± 0.08	0.15 ± 0.05	NA	NA
k	0.003 ± 0.001	0.001 ± 0.001	-	0.007 ± 0.001	0.25 ± 0.10	0.10 ± 0.05	NA	NA
m	0.277 ± 0.094	0.071 ± 0.036	-	0.387 ± 0.053	0.38 ± 0.11	0.10 ± 0.05	NA	NA
AIP	1.107 ± 0.356	0.550 ± 0.164	-	1.186 ± 0.191	0.39 ± 0.11	0.19 ± 0.06	NA	NA
WIP	0.019 ± 0.006	0.010 ± 0.003	-	0.019 ± 0.003	0.39 ± 0.11	0.22 ± 0.06	NA	NA
DFI	160.835 ± 75.157	115.647 ± 37.186	-	419.822 ± 44.426	0.23 ± 0.10	0.17 ± 0.05	NA	NA
FCR	13.207 ± 6.506	6.534 ± 3.111	NS	46.522 ± 4.117	0.20 ± 0.10	0.10 ± 0.05	NA	-
RFI	101.495 ± 43.006	44.271 ± 20.687	NS	314.588 ± 27.766	0.22 ± 0.09	0.10 ± 0.04	NA	-
RG	0.374 ± 0.184	0.378 ± 0.100	-	0.966 ± 0.106	0.22 ± 0.10	0.22 ± 0.06	NA	NA
RIG	0.375 ± 0.178	0.311 ± 0.092	-	0.905 ± 0.102	0.24 ± 0.11	0.20 ± 0.06	NA	NA
KR	0.222 ± 0.116	0.197 ± 0.060	-	0.675 ± 0.069	0.20 ± 0.10	0.18 ± 0.05	NA	NA
DOF	0.0009 ± 0.0003	NS	-	0.0054 ± 0.0003	0.14 ± 0.05	NA	NA	NA
VarF	8.865 ± 3.532	NS	-	51.352 ± 3.453	0.15 ± 0.06	NA	NA	NA

¹ELISA-G, AMDV-G based enzyme-linked immunosorbent assay test; ELISA-P, VP2 based enzyme-linked immunosorbent assay test; CIEP, counterimmunoelectrophoresis test; IAT, iodine agglutination test; BW (13–28), body weight at corresponding measurement week (13–28); HW, body weight at harvest; ADG, average daily gain; α, asymptotic weight in kg; k, second part of growth curve, in which growth rate decreases until the animal reaches the asymptotic or mature weight (α); m, shape parameter determining the position of curve inflection point; AIP, age at the inflection point; WIP, weight at the inflection point; DFI, daily feed intake; FCR, feed conversion ratio; RFI, residual feed intake; RG, residual gain; RIG, residual intake and gain; KR, Kleiber ratio; DOF, the proportion of off-feed days; VarF, variation in daily feed intake.

${\sigma }_a^2$ , additive genetic variance; ${\sigma }_m^2$, maternal genetic variance; ${\sigma }_{pe}^2$, permanent environmental variance; ${\sigma }_e^2$, residual variance.

h ² , heritability from univariate models; $c_d^2$, proportion of phenotypic variance explained by maternal genetic effects; $c_{pe}^2$, proportion of phenotypic variance explained by permanent environmental effects; r², repeatability.

NS = not significant (P > 0.05).

NA = not applicable.

-not tested.

Heritability and repeatability estimates

Heritability estimates (±SE) for all studied traits are presented in Table 3. The estimated heritabilities (±SE) of ELISA-G (0.26 ± 0.05) and IAT (0.20 ± 0.04) in the current study were lower than the estimates reported in the previous study (Hu et al., 2021) of 0.39 ± 0.06 for ELISA-G and 0.26 ± 0.05 for IAT. The larger sample size, more complete pedigree information, and more phenotypic records in this study lead to different models and variance components than the previous study (Hu et al., 2021), resulting in different estimates of heritabilities for these traits. The heritabilities of BW, growth parameters (Do et al., 2021), and feed efficiency traits (Davoudi et al., 2022) were previously reported in mink. The estimated heritabilities of BW traits (ranging from 0.23 ± 0.13 to 0.35 ± 0.11) and growth parameters (ranging from 0.17 ± 0.08 to 0.39 ± 0.11) in this study were moderate and lower than the estimates reported by Do et al. (2021). This might be due to the use of numerator relationship matrix for random maternal genetic effects in this study, compared with the identity matrix used by Do et al. (2021). The heritabilities of feed efficiency traits were estimated to be moderate and ranged from 0.20 ± 0.10 to 0.24 ± 0.11, which were similar to the estimates by Davoudi et al. (2022). The estimated heritabilities (±SE) were 0.14 ± 0.05 for DOF and 0.15 ± 0.06 for VarF, indicating the presence of small additive genetic effects for both traits. The DOF and VarF traits were new traits in mink research, and thus, no previous estimates were available for comparison. The estimated heritabilities of VarF and DOF in swine ranged from 0.08 to 0.21 and from 0.10 to 0.15, respectively, which were similar to our results for mink (Putz et al., 2019; Cheng et al., 2020, 2021).

The repeatability (±SE) of IAT was estimated to be 0.41 ± 0.04 (Table 3). Repeatability indicates the tendency of animals to maintain their performance for a certain trait over time. The moderate repeatability of IAT indicated that mink with a low IAT score would likely remain low in subsequent tests. Most farmed commercial mink will only be kept in a barn for less than nine months if they are not selected as breeders for the next season. Thus, instead of routine tests of IAT, which could cause extra stress on farmed mink, one-time IAT test results would be reliable in making selection decisions.

Correlations between Aleutian disease tests and body weight traits

The phenotypic and genetic correlations between AD tests and body weight traits are presented in Tables 4 and 5, respectively. Both ELISA-G and ELISA-P showed significant (P < 0.05) low negative phenotypic correlations with BW22, BW25, BW28, and HW (ranging from −0.08 ± 0.04 to −0.15 ± 0.04). All the phenotypic correlations were not significant (P > 0.05) between BWs at different ages and ELISA and CIEP tests, and the phenotypic correlations were slightly decreased from week 13 to week 28, which indicated that the influence caused by AD on individual BW were not getting severer with the increase of age. In the meantime, the absolute genetic correlations between AD tests and BWs at 13, 16, 19, and 22 weeks were all larger than their corresponding phenotypic correlations. The IAT showed significant (P < 0.05) low phenotypic correlations with BW13, BW16, BW28, and HW (ranged from −0.11 ± 0.04 to 0.08 ± 0.03). Although a significant loss of body weight was observed in AD-infected mink (Eklund et al., 1968), it is still unknown to what extent AD influences the BW of infected mink. Among all AD tests, only IAT showed significant (P < 0.05) moderate positive genetic correlation (0.45 ± 0.23) with BW16 and negative moderate genetic correlation (−0.47 ± 0.12) with HW. These estimates indicated that the selection of mink with lower IAT scores could decrease the BW of mink at 16 weeks of age but increase the BW of mink at harvest. No AD-specific tests (ELISA-G, ELISA-P, and CIEP) showed significant (P < 0.05) genetic correlations with body weight traits. These non-significant (P > 0.05) genetic correlations indicated that selecting a low ELISA test score or negative CIEP would not influence body weight traits. Although all the genetic correlations between BWs at different ages and AD tests were not significant, all AD tests showed the highest absolute genetic correlations with BW at 16 weeks of age compared with BWs at other ages, which indicated BW at week 16 might be the best BW indicator trait to select for favorable AD test results indirectly. Therefore, mink farmers could select mink with low anti-AMDV antibody levels based on ELISA or CIEP tests without adverse consequences on body weight.

Table 4.

Estimates of phenotypic correlations with their standard errors between Aleutian disease tests with feed efficiency, growth, and feed intake traits in mink.

Traits1	ELISA-G	ELISA-P	CIEP	IAT
BW13	−0.05 ± 0.04	−0.03 ± 0.04	−0.04 ± 0.04	0.08 ± 0.03*
BW16	−0.06 ± 0.04	−0.04 ± 0.04	−0.05 ± 0.04	0.07 ± 0.03*
BW19	−0.07 ± 0.04	−0.06 ± 0.04	−0.05 ± 0.04	0.01 ± 0.04
BW22	−0.08 ± 0.04*	−0.08 ± 0.04*	−0.06 ± 0.04	−0.02 ± 0.04
BW25	−0.08 ± 0.04*	−0.11 ± 0.04*	−0.07 ± 0.04	−0.04 ± 0.04
BW28	−0.11 ± 0.04*	−0.15 ± 0.04*	−0.07 ± 0.04	−0.07 ± 0.03*
HW	−0.12 ± 0.04*	−0.14 ± 0.04*	−0.03 ± 0.04	−0.11 ± 0.04*
ADG	−0.32 ± 0.03*	−0.15 ± 0.04*	−0.04 ± 0.04	−0.14 ± 0.04*
Α	−0.07 ± 0.03*	−0.10 ± 0.04*	−0.04 ± 0.04	−0.08 ± 0.03*
k	0.10 ± 0.04*	0.15 ± 0.04*	0.01 ± 0.04	0.13 ± 0.03*
m	0.05 ± 0.04	0.09 ± 0.04*	−0.02 ± 0.04	0.06 ± 0.04
AIP	−0.04 ± 0.04	−0.01 ± 0.04	−0.03 ± 0.04	0.02 ± 0.04
WIP	0.01 ± 0.04	−0.01 ± 0.04	−0.04 ± 0.04	0.03 ± 0.04
DFI	−0.08 ± 0.04*	−0.14 ± 0.04*	−0.05 ± 0.04	−0.03 ± 0.04
FCR	0.09 ± 0.04*	0.09 ± 0.04*	0.03 ± 0.04	0.10 ± 0.04*
RFI	0.02 ± 0.03	−0.02 ± 0.04	0.01 ± 0.04	0.03 ± 0.04
RG	−0.05 ± 0.04	−0.07 ± 0.04	0.03 ± 0.04	−0.13 ± 0.04*
RIG	−0.04 ± 0.04	−0.05 ± 0.04	0.03 ± 0.04	−0.11 ± 0.04*
KR	−0.10 ± 0.04*	−0.14 ± 0.04*	−0.02 ± 0.04	−0.15 ± 0.04*
DOF	0.02 ± 0.03	0.07 ± 0.04	−0.05 ± 0.03	0.08 ± 0.04*
VarF	0.01 ± 0.03	0.06 ± 0.04	−0.04 ± 0.03	−0.01 ± 0.03

¹ELISA-G, AMDV-G based enzyme-linked immunosorbent assay test; ELISA-P, VP2 based enzyme-linked immunosorbent assay test; CIEP, counterimmunoelectrophoresis test; IAT, iodine agglutination test; BW (13–28), body weight at corresponding measurement week (13–28); HW, body weight at harvest; ADG, average daily gain; α, asymptotic weight in kg; k, second part of growth curve, in which growth rate decreases until the animal reaches the asymptotic or mature weight (α); m, shape parameter determining the position of curve inflection point; AIP, age at the inflection point; WIP, weight at the inflection point; DFI, daily feed intake; FCR, feed conversion ratio; RFI, residual feed intake; RG, residual gain; RIG, residual intake and gain; KR, Kleiber ratio; DOF, the proportion of off-feed days; VarF, variation in daily feed intake.

^*Significant (P < 0.05).

Table 5.

Estimates of genetic correlations with their standard errors between Aleutian disease tests with feed efficiency, growth, and feed intake traits in mink

Traits1	ELISA-G	ELISA-P	CIEP	IAT
BW13	0.16 ± 0.20	0.19 ± 0.23	−0.17 ± 0.42	0.42 ± 0.27
BW16	0.23 ± 0.20	0.21 ± 0.22	−0.35 ± 0.37	0.45 ± 0.23*
BW19	0.14 ± 0.18	0.13 ± 0.18	−0.19 ± 0.35	0.21 ± 0.19
BW22	0.09 ± 0.17	0.09 ± 0.18	−0.14 ± 0.34	0.14 ± 0.18
BW25	−0.01 ± 0.18	−0.02 ± 0.19	−0.30 ± 0.32	0.01 ± 0.18
BW28	−0.06 ± 0.19	−0.16 ± 0.19	0.21 ± 0.46	−0.01 ± 0.20
HW	0.37 ± 0.26	0.01 ± 0.22	0.49 ± 0.48	−0.47 ± 0.20*
ADG	−0.37 ± 0.16*	−0.15 ± 0.20	−0.02 ± 0.40	−0.24 ± 0.18
Α	−0.01 ± 0.22	0.02 ± 0.22	0.14 ± 0.45	0.02 ± 0.22
k	0.38 ± 0.19*	0.36 ± 0.18*	−0.35 ± 0.40	0.26 ± 0.19
m	0.12 ± 0.17	0.11 ± 0.16	−0.63 ± 0.35	0.15 ± 0.17
AIP	−0.01 ± 0.16	0.03 ± 0.17	−0.58 ± 0.32	0.05 ± 0.16
WIP	0.15 ± 0.17	0.05 ± 0.17	−0.51 ± 0.31	0.25 ± 0.17
DFI	−0.13 ± 0.20	0.08 ± 0.22	0.21 ± 0.45	−0.01 ± 0.21
FCR	0.15 ± 0.25	0.09 ± 0.22	0.23 ± 0.48	0.03 ± 0.21
RFI	−0.04 ± 0.22	−0.05 ± 0.20	0.21 ± 0.41	0.02 ± 0.20
RG	0.23 ± 0.22	−0.11 ± 0.22	0.21 ± 0.42	−0.17 ± 0.21
RIG	0.20 ± 0.21	−0.08 ± 0.22	0.19 ± 0.43	−0.16 ± 0.21
KR	−0.02 ± 0.02	−0.21 ± 0.22	0.02 ± 0.48	−0.19 ± 0.20
DOF	0.22 ± 0.19	0.42 ± 0.17*	−0.02 ± 0.22	0.73 ± 0.16*
VarF	0.04 ± 0.19	0.19 ± 0.18	−0.12 ± 0.23	0.09 ± 0.21

¹ELISA-G, AMDV-G based enzyme-linked immunosorbent assay test; ELISA-P, VP2 based enzyme-linked immunosorbent assay test; CIEP, counterimmunoelectrophoresis test; IAT, iodine agglutination test; BW (13–28), body weight at corresponding measurement week (13–28); HW, body weight at harvest; ADG, average daily gain; α, asymptotic weight in kg; k, second part of growth curve, in which growth rate decreases until the animal reaches the asymptotic or mature weight (α); m, shape parameter determining the position of curve inflection point; AIP, age at the inflection point; WIP, weight at the inflection point; DFI, daily feed intake; FCR, feed conversion ratio; RFI, residual feed intake; RG, residual gain; RIG, residual intake and gain; KR, Kleiber ratio; DOF, the proportion of off-feed days; VarF, variation in daily feed intake.

^*Significant (P < 0.05).

This study is the first study to estimate the phenotypic and genetic correlations between AD tests and body weight traits; thus, no estimates were available in the literature for direct comparison. The genetic correlations between disease traits and body weight traits were estimated in other livestock species. For example, in feedlot cattle, the bovine respiratory disease had a low genetic correlation (0.04) with hot carcass weight (Snowder et al., 2007). In dairy cattle, the genetic correlations between infectious diseases (e.g. endotoxemia, enteritis, foot rot, mastitis, metritis, and pneumonia) and BW on different days in milk production ranged from −0.05 to −0.81 (Frigo et al., 2010). In sheep, the genetic correlations between body weight and fecal egg count (used to measure the resistance to nematode parasites) were low-to-moderate negative ranging from −0.18 to −0.26 (Eady et al., 1998). In chicken, the antibody response against the Newcastle disease virus vaccine was estimated to have low genetic correlations (range from −0.04 to −0.08) with BW at 8, 12, 16, and 20 weeks of age (Lwelamira et al., 2009).

Correlations between Aleutian disease tests and growth parameters

The phenotypic and genetic correlations between AD tests and growth traits are shown in Tables 4 and 5, respectively. IAT showed significant (P < 0.05) low negative phenotypic correlation (−0.14 ± 0.04) and non-significant (P > 0.05) negative genetic correlation (−0.24 ± 0.18) with ADG. These results revealed that higher serum gamma globulin level decreases the average daily weight gain of mink, which could be mainly due to the environmental effects. The IAT is a non-AD-specific test and is designed to diagnose AD by detecting hypergammaglobulinemia, which could be caused by other environmental factors such as stress, temperature, and other diseases (Delves and Roitt, 1998); thus, the phenotypic correlation between IAT and ADG mainly could be attributed to the environmental effects. Among all AD test traits, only ELISA-G showed significant (P < 0.05) moderate negative phenotypic (−0.32 ± 0.03) and genetic (−0.37 ± 0.16) correlations with ADG. This indicated that selection for a low ELISA-G score would increase the ADG. Meanwhile, ADG could be applied as a low-stress and economic indicator to indirectly select mink with low ELISA-G scores. The non-significant (P > 0.05) genetic correlations between ADG with ELISA-P, CIEP, and IAT indicated that the selection of mink with low ELISA-P, IAT, or negative CIEP results would not cause significant influences on mink growth. The genetic correlation between AD and ADG was not estimated in previous research in mink. The genetic correlations between disease traits and ADG were widely estimated in other livestock species, and the genetic correlations were usually low or not significant, which were in agreement with our results. For example, several studies reported that cattle with bovine respiratory disease generally had lower ADG than healthy individuals (Bateman et al., 1990; Gardner et al., 1999; Snowder et al., 2006), but the genetic correlations were not significant (Snowder et al., 2007). In addition, the parasite burden traits in cattle, including tick infestation, gastrointestinal nematodes infection, and Eimeria spp. infection, were estimated to have low genetic correlations (from −0.22 to 0.12) with ADG (Biegelmeyer et al., 2015; Ribeiro et al., 2021). In swine, several studies reported that lung lesions caused by Mycoplasma pneumonia could significantly reduce ADG (Okada et al., 1999; Wilson et al., 2012) and had a low (0.002) genetic correlation with ADG (Okamura et al., 2016). In sheep, the ewe fecal egg counts, which were used to assess the severity of Teladorsagia circumcincta and Haemonchus contortus infection, was estimated to have a low genetic correlation (−0.15) with ADG (Bouix et al., 1998).

Both ELISA-G and ELISA-P showed significant (P < 0.05) moderate positive genetic correlations (0.38 ± 0.19 and 0.36 ± 0.18, respectively) with parameter k, which indicated that selection of mink with lower anti-AMDV antibody levels could decrease the maturation rate of mink. The smaller magnitude of the k parameter will lead to later mature growth (Do et al., 2021). The relationship between later mature growth and other economically important traits (such as fur size, fur quality, and reproduction) has not been studied, but mink with lower k tends to have larger mature body weight, and thus may lead to larger harvest length (Do et al., 2021). The previous study (Hu et al., 2021) estimated significant (P < 0.05) moderate negative phenotypic (−0.30 ± 0.06) and genetic (−0.45 ± 0.16) correlations between ELISA-G and harvest length. Therefore, the significant positive correlations between ELISA tests and the k growth parameter could be treated as favorable correlations, but further investigation is needed. Although most of the genetic correlations between growth parameters and AD tests were not significant, the absolute genetic correlations between AD tests with k, m, and WIP were larger than their corresponding absolute phenotypic correlations. Both ELISA-G and ELISA-P showed significant (P < 0.05) low negative phenotypic correlations with α (−0.07 ± 0.03 and −0.10 ± 0.04, respectively), but non-significant (P > 0.05) genetic correlations with α (−0.01 ± 0.22 and 0.02 ± 0.22, respectively). These estimates revealed that higher anti-AMDV antibody level might decrease the mature body weight of mink, which could be mainly due to non-genetic effects. Although it has been observed that higher immune response could cause adverse influences on growth in livestock, the underlying mechanism is still unclear (Doeschl-Wilson et al., 2009; van der Most et al., 2011; Hu et al., 2020). Energy allocation could be one of the reasons leading to the unfavorable relationship between animal growth and immune response because the resources spent on immune response were no longer available for growth and other energy-consuming processes, such as locomotion and thermoregulation (van der Most et al., 2011).

Correlations between Aleutian disease tests and feed efficiency traits

Phenotypic and genetic correlations between AD tests and feed efficiency traits are shown in Tables 4 and 5, respectively. The phenotypic correlations between AD tests and feed efficiency traits were either low (ranged from −0.15 ± 0.04 to 0.09 ± 0.04) or not significant (P > 0.05). None of the AD test traits showed a significant (P < 0.05) genetic correlation with feed efficiency traits. Although CIEP did not show significant genetic correlations with any feed efficiency traits (DFI, FCR, RFI, RG, and RIG), their magnitudes were very close or consistent (ranged from 0.19 ± 0.43 to 0.23 ± 0.48). The genetic correlation between AD and feed efficiency was not estimated in previous studies in mink for direct comparison. The genetic correlations between disease and feed efficiency traits were rarely estimated, even in the main livestock species. In poultry, the ratio of the weight of right ventricle to the total ventricles, which is used to measure the ascites syndrome, was estimated to have low negative genetic correlations (−0.17 to −0.19) with feed intake between 23 and 48 days of age and residual feed intake (Pakdel et al., 2005). Meanwhile, a moderate positive (0.38) genetic correlation between total ventricles and feed efficiency, which was defined as the ratio between body weight gain and feed intake during the experiment, was also detected (Pakdel et al., 2005). In swine, osteochondrosis lesions were estimated to have low-to-moderate genetic correlations with FCR, ranging from −0.23 to 0.40 (Kadarmideen et al., 2004). In cattle, the mastitis incidence in Danish Friesian cattle was estimated to have a significant (P < 0.05) low negative genetic correlation (−0.15) with feed efficiency, and the mastitis incidence in Danish Jersey cattle was estimated to have a significant (P < 0.05) high negative genetic correlation (−0.79) with feed efficiency (Wassmuth et al., 2000). In addition, ketosis incidence in Danish Jersey cattle was estimated to have a significant (P < 0.05) moderate negative genetic correlation (−0.37) with feed efficiency (Wassmuth et al., 2000).

The non-significant (P > 0.05) genetic correlations between AD tests and feed efficiency traits indicated that the selection of mink with favorable AD test results (low ELISA and IAT score and negative CIEP) would not cause a significant influence on the feed efficiency of mink. Feed is the largest cost, up to 70% of the total, in mink farming (Berg and Lohi, 1992; Sørensen et al., 2003), and AD is the most severe disease that causes tremendous economic losses to mink farmers. Together, these two can significantly influence the financial return for mink farmers. The genetic correlations between AD test and feed efficiency estimated in this study indicated that mink breeders could select mink with higher feed efficiency and AD resilience simultaneously through genetic selection without adverse influences on farmers’ profitability. This would significantly contribute to producing efficient mink, which are resilient to AD and feed efficient, to minimize the economic losses caused by AD, reduce environmental impacts from mink production, and improve mink profitability.

Correlations between Aleutian disease tests and feed-intake-related traits

Phenotypic and genetic correlations between AD tests and feed intake traits derived from DFI data are shown in Tables 4 and 5, respectively. Among all AD tests, only IAT showed a significant (P < 0.05) low positive phenotypic (0.08 ± 0.04) correlation with DOF. The ELISA-P showed a significant (P < 0.05) moderate positive genetic correlation (0.42 ± 0.17) with DOF, and IAT showed a significant (P < 0.05) high positive genetic correlation (0.73 ± 0.16) with DOF. These estimates indicated that selection of mink with lower ELISA-P or IAT scores could reduce the off-feed days, and thus reduce the adverse influences caused by AD on mink appetite in AD positive farms. No significant phenotypic or genetic correlations between AD tests and VarF were estimated, indicating the selection of AD tests on mink would not cause significant influences on the consistency of feed intake.

Except for this study, the genetic correlations between disease traits with DOF and VarF have only been estimated in swine because they are novel resilient traits recently used in swine (Putz et al., 2019). DOF was estimated to have moderate-to-high genetic correlations with the number of health treatments (range from 0.50 to 0.85) and mortality (0.30–0.80). The VarF was estimated to have moderate-to-high genetic correlations with the number of health treatments (range from −0.30 to 0.60) and mortality (0.37 to 0.85) in a natural disease challenge swine farm (Putz et al., 2019; Cheng et al., 2020).

Aleutian disease resilience indicator traits based on Aleutian disease tests

In recent years, numerous studies have suggested that selection of resilience should be used as a practical method to cope with diseases in livestock industries (Berghof et al., 2019a; Harlizius et al., 2020; Hu et al., 2020; Iung et al., 2020; Knap and Doeschl-Wilson, 2020). Resilience is defined as an animal’s ability to maintain its performance under pathogen exposure (Albers et al., 1987; Bisset and Morris, 1996). A reliable and practical disease resilience indicator trait genetically correlated with health traits is essential for the genetic selection of healthy animals (Mulder and Rashidi, 2017; Berghof et al., 2019a). Body weight, growth, and feed intake traits have been suggested as indicators for specific disease resilience traits or general resilience traits in farm animals (Colditz and Hine, 2016; Berghof et al., 2019a). For example, in layer chickens, the standardized body weight deviations were suggested as general resilience traits (Berghof et al., 2019b). In swine, growth rate was suggested as a disease resilience trait for porcine reproductive and respiratory syndrome (Hess et al., 2018; Knap and Doeschl-Wilson, 2020), and traits derived from feed intake data were suggested as general disease resilience traits (Putz et al., 2019; Cheng et al., 2020, 2021). In mink, AD has been reported to cause adverse influences on body weight, growth, and appetite (Eklund et al., 1968; Porter et al., 1982; Jensen et al., 2016); thus, the studied traits of BW, growth, feed efficiency, and feed intake traits, could be treated as AD resilience traits in the AD positive farms.

Among all AD tests, ELISA-G may be the most practical and reliable test to be employed as an indicator for selecting AD-resilient mink. The CIEP did not show any significant genetic correlations with other AD resilience traits in this study, and together with its low heritability (0.09 ± 0.07), it may not be a good indicator for AD resilience. Due to the low test price and easy test procedure of IAT, together with assessing the production traits, has been used by mink farms to select for AD-resilience in North America (Farid and Ferns, 2011). Although IAT was estimated to have moderate heritability (0.20 ± 0.04) and repeatability (0.41 ± 0.04) and showed significant favorable moderate genetic correlations with ELISA tests (Hu et al., 2021), DOF (0.73 ± 0.16), and HW(−0.47 ± 0.20), the potential adverse effects on pelt nap length (the length of the guard hair protruding from the underwool) caused by selecting for lower IAT scores cannot be ignored (Hu et al., 2021). Therefore, IAT may not be the best indicator trait for selecting AD-resilient mink. Although ELISA-P showed the highest heritability (0.59 ± 0.07) among all AD test traits with significant favorable genetic correlations with k and DOF, the unknown repeatability may require mink farmers to have multiple ELISA-P tests on the same individual, which would increase the expense of AD tests. Meanwhile, compared with ELISA-G, ELISA-P did not show significant (P < 0.05) favorable genetic correlations with female reproductive performance traits, which are important AD resilient traits (Hu et al., 2021). Thus, further investigation is required before using ELISA-P in selecting AD-resilient mink. The ELISA-G was estimated to have moderate heritability (0.26 ± 0.05) and high repeatability (0.52 ± 0.04), indicating that the one-time ELISA-G test could be a reliable indicator for selecting mink with low anti-AMDV antibody level to reduce the formation of infectious virus-antibody complexes. Meanwhile, ELISA-G showed significant (P < 0.05) favorable genetic correlations with ADG and k growth parameter and non-significant (P > 0.05) genetic correlations with other growth parameters, BW, feed efficiency, DOF, and VarF. All these estimated genetic parameters for ELISA-G indicated that selection of mink with lower AMDV-G ELISA could improve the average daily gain and bring benefits to mature weight without causing adverse influences on body weight, feed efficiency, off-feed days, and consistency of feed intake. Meanwhile, the previous study (Hu et al., 2021) showed that selection for low ELISA-G test results could also increase both female reproductive performance traits and harvest length and decrease the extent of anemia without influencing pelt quality in AD positive farms. Therefore, ELISA-G could be the most practical and reliable indicator in genetic selection for selecting AD-resilient mink in AD positive mink farms compared to other AD tests.

Conclusions

Finding a practical solution to control AD is one of the top priorities for the mink industry due to the tremendous financial losses caused by AD. Genetic selection for AD-resilient mink would provide a great opportunity to reduce the adverse influences caused by AD in mink farming. This research was a further supplement and exploration of the potential of AD tests in genetic selection of AD-resilient mink by involving more economically important or AD-resilient traits, thus providing a more reliable and comprehensive view of employing AD tests in AD-resilient mink selection. The estimated genetic parameters for AD tests in this study further demonstrated that ELISA-G was the most reliable and practical indicator trait among all AD tests that could be applied in the genetic selection of AD-resilient mink. However, unlike genetic studies on other main livestock species such as cattle, swine, and poultry, which usually had larger sample sizes due to the longer data collection histories, wider data collection ranges, and more complete data-sharing systems, the sample size used in this study was relatively smaller. Thus, further studies with larger sample sizes and more populations/farms are required to validate the current findings. In addition, future genomic studies, such as genome-wide association studies, linkage map and QTL fine mapping, could be applied to better understand the biological mechanisms behind the correlations between AD tests and AD-resilient or other economically important traits in mink. In the absence of any other AD-resilient-related indicator trait, ELISA-G could be applied as the most useful and practical indicator trait to select AD-resilient mink in AD positive farms.

Glossary

Abbreviations

AD

Aleutian disease

ADFI

average daily feed intake

ADG

average daily gain

AIP

age at the inflection point

AMDV

Aleutian mink disease virus

AMDV-G

in vitro cultured Aleutian mink disease virus antigen

BW

body weight

CCFAR

Canadian Centre for Fur Animal Research

CIEP

counterimmunoelectrophoresis

DFI

daily feed intake

DOF

proportion of off-feed days based on feed intake

ELISA

enzyme-linked immunosorbent assay

ELISA-G

in vitro cultured Aleutian mink disease virus antigen-based enzyme-linked immunosorbent assay test

ELISA-P

capsid protein of Aleutian mink disease virus-based enzyme-linked immunosorbent assay test

FCR

feed conversion ratio

h ²

heritability

HW

body weight at harvest

IAT

iodine agglutination test

k

maturation rate

KR

Kleiber ratio

m

inflection parameter

NA

not applicable

NS

not significant

NT

not tested

r ²

repeatability

RFI

residual feed intake

RG

residual gain

RIG

residual intake and gain

VarF

day-to-day variation in feed intake

VP2

capsid protein of Aleutian mink disease virus

WIP

weight at the inflection point

α

asymptotic weight

β

parameter characterizing the first part of growth before the inflection point

${\sigma }_a^2$

additive genetic variance

${\sigma }_m^2$

maternal genetic variance

${\sigma }_{pe}^2$

permanent environmental variance

${\sigma }_e^2$

residual variance

Conflict of Interest Statement

The authors declare no conflict of interest. The funders had no role in the design of the study, data collection, data analyses, interpretation, writing of the manuscript, and the decision to publish the results.