Individual variation in sow aggressive behavior and its relationship with sow welfare

Estimation of gestating sows' welfare status based on machine learning methods and behavioral data

Estimating the welfare status at an individual level on the farm is a current issue to improve livestock animal monitoring. New technologies showed opportunities to analyze livestock behavior with machine learning and sensors. The aim of the study was to estimate some components of the welfare status of gestating sows based on machine learning methods and behavioral data. The dataset used was a combination of individual and group measures of behavior (activity, social and feeding behaviors). A clustering method was used to estimate the welfare status of 69 sows (housed in four groups) during different periods (sum of 2 days per week) of gestation (between 6 and 10 periods, depending on the group). Three clusters were identified and labelled (scapegoat, gentle and aggressive). Environmental conditions and the sows' health influenced the proportion of sows in each cluster, contrary to the characteristics of the sow (age, body weight or body condition). The results also confirmed the importance of group behavior on the welfare of each individual. A decision tree was learned and used to classify the sows into the three categories of welfare issued from the clustering step. This classification relied on data obtained from an automatic feeder and automated video analysis, achieving an accuracy rate exceeding 72%. This study showed the potential of an automatic decision support system to categorize welfare based on the behavior of each gestating sow and the group of sows.

Social Status Differentially Affects Behavioral and Immunological Outcomes of Group-Kept Sows Fed Different Dietary Fiber Using Different Length Feeding Barriers

Social conflict is inevitable among group-housed sows and may contribute to poorer welfare among those sows experiencing more social stress. The degree of individual welfare is associated with social position within the group. Therefore, this study examined the effects of social status on behavior, immune, endocrine, and productivity of group-housed pregnant sows fed a diet supplemented with 30% wheat middlings and 15% soybean hulls (MID-SH) or 30% distillers dried grains with solubles and 30% corn germ meal (DDGS-GM) and in pens with individual feeding places made from short (58.4 cm) or long (203.2 cm) barriers. A 2 × 2 factorial design resulted in 4 experimental treatment groups (n = 9 sows/diet-length-block combination): (1) MID-SHshort; (2) MID-SHlong; (3) DDGS-GMshort; (4) DDGS-GMlong. Groups of sows equally representing all diet-length combinations across 4 blocks (n = 36 sows/block) were subjected to a feeding competition test to identify highest (dominant) and lowest (subordinate) ranked sows within each group resulting in 64 sows (n = 16 sows/treatment; n = 32 sows/social status). Data revealed 2- and 3-way interactive effects on aggressive behavior (P < 0.005), postural (P < 0.01), oral (P < 0.0001), and eating (P < 0.005) behaviors, sow mean body weights and gains (P < 0.05) and litter weaning weights (P < 0.05), especially among subordinates in pens with long barriers. Subordinates in pens with long barriers received 21% less aggression and were 73% less likely to be displaced than subordinates in pens with short ones (P < 0.0001). Dietary treatment also influenced some of these measures among the subordinates in pens with long barriers. For example, subordinates in DDGS-GMlong received 64 and 67% less aggression than subordinates in DDGS-GMshort and MID-SHshort (P < 0.005). Eat bouts were greatest among subordinates in MID-SHlong, and sitting and sham-chewing were less. However, those in DDGS-GMlong spent less time standing and laying, and their litters were 15.28 kg heavier (P = 0.01), but overall subordinates fed DDGS-GM diet were lightest and gained less total body weight than those fed MID-SH (P < 0.05). Other measures such as neutrophil-to-lymphocyte ratio were elevated among dominants in MID-SHlong (P < 0.05); whereas, cortisol (P = 0.06) was lowest and glucose (P = 0.09) highest for subordinates in DDGS-GMlong. These data imply that subordinates benefited from being housed in pens with long barriers, but the type of dietary fiber consumed differentially influenced behavioral budget and several sow- or litter-related traits among subordinates in pens with long barriers. In contrast, the subordinates in pens with short barriers had poorer welfare regardless of diet. Collectively, these data imply that social status is a crucial factor contributing to variation in individual well-being among group-housed sows and that sows of different social positions within a group may evoke different biological responses in an attempt to cope.

Social status and previous experience in the group as predictors of welfare of sows housed in large semi-static groups

Mixing gestating sows implies hierarchy formation and has detrimental consequences on welfare. The effects of social stress on the most vulnerable individuals may be underestimated and it is therefore important to evaluate welfare between individuals within groups. This study aimed at investigating the impact of social status and previous experience in the group on well-being of sows housed in large semi-static groups. We assessed aggression (d0 (mixing), d2, d27, d29), body lesions (d1, d26, d84) and feeding order on 20 groups of 46–91 animals. Social status was based on the proportion of fights won during a 6-hr observation period between d0 and d2. Dominants (29%) were those who won more fights than they lost, Subdominants (25%) won fewer fights than they lost, Losers (23%) never won any fight in which they were involved while Avoiders (23%) were never involved in fights. Resident sows (70%) were already present in the group in the previous gestation while New sows (30%) were newly introduced at mixing. Subdominants and Dominants were highly involved in fights around mixing but this was more detrimental for Subdominants than Dominants, Losers and Avoiders since they had the highest body lesion scores at mixing. Avoiders received less non-reciprocal agonistic acts than Losers on d2 (P = 0.0001) and had the lowest body lesion scores after mixing. However, Avoiders and Losers were more at risk in the long-term since they had the highest body lesions scores at d26 and d84. They were followed by Subdominants and then Dominants. New sows fought more (P<0.0001), tended to be involved in longer fights (P = 0.075) around mixing and had more body lesions throughout gestation than Resident sows. Feeding order from one-month post-mixing was influenced both by the previous experience in the group and social status (P<0.0001). New sows, especially with a low social status, are more vulnerable throughout gestation and could serve as indicators of non-optimal conditions.

Effects of supplementing sow diets during late gestation with Pennisetum purpureum on antioxidant indices, immune parameters and faecal microbiota

The purpose of this study was to investigate the effects of adding Pennisetum purpureum (P. purpureum, also known as Napier grass or elephant grass) to the diets of late gestation on the antioxidant indexes, immune indexes and faecal microbiota of sows. At the 90 days of gestation, 300 healthy sows were randomly divided into three groups, and they received the basic commercial diet or added 5% P. purpureum and 10% P. purpureum, respectively. The experiment started from 90 days of gestation to parturition. The results showed that the total antioxidant capacity, immunoglobulins and serum equol concentrations of sows on 100 days of gestation and at parturition increased linearly (p < .05) with the increase of the content of P. purpureum in the gestation diet. The 5% P. purpureum increased the relative abundance of Bacteroidetes (p = .027) and Actinobacteria (p < .001) at phylum level, Coriobacteriaceae (p < .001) at family level and Prevotellaceae_UCG_001 (p = .004) at genus level, and decreased the relative abundance of Escherichia_Shigella (p < .001) at genus level. In summary, this study shows that the additive of P. purpureum can increase the concentration of serum equol, improve the antioxidant capacity and immune function of sow in late gestation. In addition, the additive of 5% P. purpureum in the diet might change the composition of intestinal microbiota of sows, particularly the relative abundance of Coriobacteriaceae (p < .001) increased.

Mixing aggression intensity is associated with age at first service and floor type during gestation, with implications for sow reproductive performance

Aggression resulting from mixing to establish a dominance hierarchy is a major welfare concern for group-housed sows. The associated stress can negatively impact aspects of reproductive performance. Objectives of this study were to investigate associations between 1) age at first service (AFS) and mixing aggression intensity in first parity sows, 2) mixing aggression intensity and reproductive performance within and between parity one and parity two, and 3) mixing aggression intensity, floor type during gestation and reproductive performance. Gilts (n =160, hereafter referred to as sows) were mixed into stable groups of eight unfamiliar individuals approximately 4 days after artificial insemination, housed on fully slatted concrete (CON; n =80) floor uncovered or covered with rubber slat mats (RUB; n =80), and followed through two parities. Skin lesions (SLMIX; a proxy for the intensity of mixing aggression), were scored post mixing in each parity according to severity (0=no lesions to 5=severe lesions) on five body regions (ear, neck, hindquarter, rump, and belly) on the left and right sides, and at the tail/anogenital region. Total SLMIX score was calculated for each sow. Data on reproductive performance traits were acquired retrospectively from farm records for both parities. Two analyses were performed: 1) data from each parity were analysed separately and 2) SLMIX score in parity one was used to predict reproductive performance in parity two. Lower AFS was associated with a lower SLMIX score in parity one (P =0.031). There was no association between SLMIX score and reproductive performance in parity one, while sows with higher SLMIX score in parity two had a higher proportion of piglets dead during lactation (P =0.027) and a longer cycle length (P =0.003) in parity two. Sows with higher SLMIX scores in parity one had more non-productive days (P <0.001) in parity two. Concrete sows had a higher SLMIX score than RUB sows in parity one (P =0.015), but not in parity two. In addition, CON sows had a higher proportion of piglets born dead (P =0.013) compared with RUB sows in parity two. Mixing aggression has a negative influence on reproductive performance within parities, and it may also have a long-term negative carry-over effect on reproductive performance in subsequent parities. Serving gilts at younger ages could help to minimize the intensity of aggression at mixing, while housing on rubber flooring has beneficial implications for their reproductive performance.

The Role of Genetic Selection on Agonistic Behavior and Welfare of Gestating Sows Housed in Large Semi-Static Groups

Simple Summary

Group-housing of gestating sows is becoming increasingly common worldwide as it offers the sows the opportunity to exercise, display exploratory behaviors, and develop social relationships. Despite its advantages, group-housing as it stands in modern industries also presents several welfare issues such as overt aggression between pen-mates and resulting stress and injuries. To date, breeding companies often largely focused their efforts on genetic selection based on individual production characteristics (e.g., litter size, piglets’ growth, and meat quality) and traditionally ignored the social behaviors and the ability to establish a dominance hierarchy without exacerbated aggression. Hence, the extent to which agonistic behavior differs according to new genetic lines is unknown. The objective of this study was to compare and investigate the influence of two genetic lines on the welfare and performance of sows housed in large semi-static groups (up to 91 animals). While the first genetic line was more aggressive toward pen-mates during gestation, the second had piglets with a lower robustness and survivability. This study raises the difficulty of finding an optimal genetic line, including both positive welfare and productivity outcomes, and points to the urgent need of considering social aspects when developing genetic lines for group-housing.

Abstract

Confinement of gestating sows is becoming banished in favor of group-housing in countries worldwide, forcing breeding companies to develop genetic lines adapted for social living. This study aimed at assessing the influence of two genetic lines selected for high performance (HP1, HP2, derived from Landrace × Yorkshire) on welfare and reproductive performance of sows housed in large semi-static groups (20 groups of 46–91 animals) across several parities. To address this, agonistic behaviors were recorded on d0, d2, d27, and d29 post-mixing while body lesions were scored on d1, d26, and d84. Sows’ individual and reproductive performances were also recorded. HP2 sows were more aggressive than HP1 sows since they fought (p = 0.028) and bullied (p = 0.0009) pen-mates more frequently on d0–d2. HP2 sows had more total body lesions throughout gestation than HP1 sows at higher parities (p < 0.0001). Regarding reproductive performance, HP2 sows lost less piglets (p < 0.0001) and tended to wean more piglets (p = 0.067) than HP1 sows. In conclusion, while HP2 sows were the most aggressive, HP1 sows had piglets with lower survivability, which raises ethical issues in both cases and points to the need of considering social aspects when developing genetic lines for group-housing.

Pregnancy-Associated Alterations of Peripheral Blood Immune Cell Numbers in Domestic Sows Are Modified by Social Rank

During pregnancy, the maternal immune system is characterized by a shift from adaptive to innate immune functions. Besides, the immune system can be influenced by social rank. Detailed knowledge of pregnancy-associated immune changes and of the interplay of rank-associated and gestation-induced immunomodulations is still fragmentary in sows. This study investigates both the numbers of various blood leukocyte subpopulations during pregnancy and the influence of social rank position on progressing pregnancy-associated alterations in group-housed sows. Sows were classified as low (LR), middle (MR), or high-ranking (HR). Five blood samples were collected from each of the 35 sows throughout pregnancy to evaluate the distribution of blood lymphocyte subpopulations and plasma cortisol concentrations. The numbers of T, natural killer (NK), and B cells, cytotoxic T cells (CTL), and CD8⁺ γδ- T cells decreased during the last trimester of pregnancy, while neutrophils and plasma cortisol concentration increased before parturition. Social rank revealed different effects on B cells and monocytes with MR sows showing higher numbers than LR sows. Plasma cortisol concentrations also tended to be higher in MR sows as compared to LR sows. In conclusion, sows show pregnancy-associated alterations in the immune system, which are influenced by social rank, as middle-ranking sows in particular display signs of stress-induced immunomodulations.

Body condition, live weight and success in agonistic encounters in mixed parity groups of sows during gestation

Group housing of gestating sows benefits their welfare by allowing them freedom of movement and the opportunity for social interaction. However, social life could also bring disadvantages for individuals who receive direct aggression or are displaced from the feeder. The aim of this study was to investigate associations between social behaviour, body condition and live weight. Gestating sows (n=298) were investigated on a commercial farm. Sows were housed in mixed parity groups where two single space, ad libitum trough feeders served 12 animals. Sows were weighed, body condition scored and had their back fat layer measured at mixing, 4 weeks after insemination and again before farrowing. Social status was estimated based on the numbers of won and lost agonistic interactions at mixing and at the end of gestation. In addition, tear staining was scored before the farrowing and reproductive performance data were collected. With the aid of video recordings, 100 to 150 interactions per group were observed. Winning percentage at mixing and at the end of gestation were associated (P&lt;0.05) and appeared relatively stable within individuals. Tear staining scores and litter sizes were not associated with winning percentage at the end of gestation. However, live weight, relative weight, body condition and back fat thickness were associated with winning percentage (P&lt;0.05), giving heavier animals an advantage. Low winning percentage related to lower live weight gain, probably due to poorer success in competition for feed. Live weight within a mixed parity group could be used as a proxy measure for social status. Sows with low body condition score and submissive sows might need special attention with regard to group dynamics and housing to alleviate the effects of competition in group housing.

Floor Feeding Sows Their Daily Allocation over Multiple Drops per Day Does Not Result in More Equitable Feeding Opportunities in Later Drops

This research studied whether floor feeding group-housed sows their daily allocation over multiple feed drops per day provides more equitable feeding opportunities in later drops. Over four time replicates, 275 sows were mixed into groups of 10 for both their first and second gestations (200 sows/gestation, 126 sows observed in both gestations). The feeding behavior of individual sows was recorded for 10 min following each of four feed drops per day (0730, 0900, 1100, 1500 h) on days 2, 9 and 51 post-mixing. The location of feeding sows (i.e., feeding in areas associated with high, reduced or little/no food availability) was also recorded. Sow aggressive behavior on day 2 was used to classify sows as dominant (D), subdominant (SD), or submissive (SM). Dominant sows spent the most time feeding in areas of high-food availability (gestation 1, p < 0.001; gestation 2, p = 0.023); SD sows fed more frequently than D sows from areas of reduced food availability (gestation 1, p = 0.001; gestation 2, p = 0.025); and SM sows performed more feeding behavior in areas of little/no food availability (gestation 1, p < 0.001; gestation 2, p < 0.001). These relationships did not change over feed drops or days in either gestation (p > 0.05). Further research on the management and design of floor feeding systems is required, with a particular emphasis on increasing accessibility to sows that avoid the feeding area.

Forming groups of aggressive sows based on a predictive test of aggression does not affect overall sow aggression or welfare

This experiment examined the effects of group composition on sow aggressive behaviour and welfare. Over 6 time replicates, 360 sows (parity 1-6) were mixed into groups (10 sows per pen, 1.8 m²/sow) composed of animals that were predicted to be aggressive (n = 18 pens) or groups composed of animals that were randomly selected (n = 18 pens). Predicted aggressive sows were selected based on a model-pig test that has been shown to be related to the aggressive behaviour of parity 2 sows when subsequently mixed in groups. Measurements were taken on aggression delivered post-mixing, and aggression delivered around feeding, fresh skin injuries and plasma cortisol concentrations at days 2 and 24 post-mixing. Live weight gain, litter size (born alive, total born, stillborn piglets), and farrowing rate were also recorded. Manipulating the group composition based on predicted sow aggressiveness had no effect (P > 0.05) on sow aggression delivered at mixing or around feeding, fresh injuries, cortisol, weight gain from day 2 to day 24, farrowing rate, or litter size. The lack of treatment effects in the present experiment could be attributed to (1) a failure of the model-pig test to predict aggression in older sows in groups, or (2) the dependence of the expression of the aggressive phenotype on factors such as social experience and characteristics (e.g., physical size and aggressive phenotype) of pen mates. This research draws attention to the intrinsic difficulties associated with predicting behaviour across contexts, particularly when the behaviour is highly dependent on interactions with conspecifics, and highlights the social complexities involved in the presentation of a behavioural phenotype.

Social status and housing factors affect reproductive performance of pregnant sows in groups

Group-housing systems for pregnant sows are considered a welfare-promoting alternative to the individual stall. A major concern associated with pregnant sows housed in group pens is increased aggression at mixing and at feeding, which may cause chronic stress among some of the sows in the group due to low feed intake and social stress. Prolonged activation of the stress axis, based on elevated cortisol levels, may inhibit or impair reproductive success via disruption of the reproductive axis. Mixing sows into groups shortly after insemination evokes a stress response, which may affect fertilization and implantation due to sustained, elevated cortisol levels that disrupt reproductive processes. Yet, most studies reported minimal effects of group housing sows during pregnancy on reproduction or cortisol-related stress response. Differences between housing systems-in terms of group size, floor-space allowance, feeding system, and genetics-could account for these unexpected results. Indeed, interrupted feed intake, especially in early pregnancy, and sustained aggression in late pregnancy are two unfavorable social stresses that deserve special attention in order to achieve good reproductive performance. Unfortunately, most studies do not consider other factors, such as social rank and parity, which may interactively affect reproductive success and aggressive behavior of sows, especially in group-pen systems.

The short-term behavioural response of sows, but not gilts, to a social stimulus is related to sow aggressiveness in groups

This study examined relationships between the behavioural response of pregnant gilts (n=200, gestation 1) and sows (n=200, gestation 2) to a live, similarly-aged female pig (unfamiliar pig test, UPT) and to a fibre-glass model pig (model pig test, MPT), and aggressive behaviour on the day after mixing (day 2). Sows with a short latency to make contact with an unfamiliar sow in the UPT were more likely to deliver high levels of aggression at day 2 of gestation 2 (P=0.005), but this relationship was stronger when a model pig was used (P<0.001). Similarly, sows with a long duration of tactile contact with the model pig in the MPT were more likely to deliver high levels of aggression at day 2 of gestation 2 (P=0.015), but this relationship was weaker than that between aggression and the latency to contact the model pig. When the terms the latency to contact the unfamiliar pig in the UPT and the model pig in the MPT, as well as the duration of contact with the model pig in the MPT, were included in an overall model of aggression at day 2 of gestation 2, behaviour towards the unfamiliar pig became not statistically significant (P>0.05). A strong relationship was not apparent with gilts (P>0.05). Thus, the socially inexperienced pig may not be an ideal model for sow behaviour. This study indicates that sows with a short latency to contact a model pig are more likely to be aggressive when mixed into groups.