Development and validation of a fully automated 2-dimensional imaging system generating body condition scores for dairy cows using machine learning

Monitoring body condition score (BCS) is a useful management tool to estimate the energy reserves of an individual cow or a group of cows. The aim of this study was to develop and evaluate the performance of a fully automated 2-dimensional imaging system using a machine learning algorithm to generate real-time BCS for dairy cows. Two separate datasets were used for training and testing. The training dataset included 34,150 manual BCS (MAN_BCS) assigned by 5 experienced veterinarians during 35 visits at 7 dairy farms. Ordinal regression methods and deep learning architecture were used when developing the algorithm. Subsequently, the testing dataset was used to evaluate the developed BCS prediction algorithm on 4 of the participating farms. An experienced human assessor (HA1) visited these farms and performed 8 whole-milking-herd BCS sessions. Each farm was visited twice, allowing for 30 d (±2 d) to pass between visits. The MAN_BCS assigned by HA1 were considered the ground truth data. At the end of the validation study, MAN_BCS were merged with the stored automated BCS (AI_BCS), resulting in a testing dataset of 9,657 single BCS. A total of 3,817 cows in the testing dataset were scored twice 30 d (±2 d) apart, and the change in their BCS (ΔBCS) was calculated. A subset of cows at one farm were scored twice on consecutive days to evaluate the within-observer agreement of both the human assessor and the system. The manual BCS of 2 more assessors (HA2 and HA3) were used to assess the interobserver agreement between humans. Finally, we also collected ultrasound measurements of backfat thickness (BFT) from 111 randomly selected cows with available MAN_BCS and AI_BCS. Using the testing dataset, intra- and interobserver agreement for single BCS and ΔBCS were estimated by calculating the simple percentage agreement (PA) at 3 error levels and the weighted kappa (κw) for the exact agreement. A Bland-Altman plot was constructed to visualize the systematic and proportional bias. The association between MAN_BCS and AI_BCS and the BFT was assessed with Passing-Bablok regressions. The system had an almost perfect repeatability with a κw of 0.99. The agreement between MAN_BCS and AI_BCS was substantial, with an overall κw of 0.69. The overall PA at the exact, ± 0.25-unit, and ± 0.50-unit BCS error range between MAN_BCS and AI_BCS was 44.4%, 84.6%, and 94.8%, respectively, and greater than the PA obtained between HA1 and HA3. The Bland-Altman plot revealed a minimal systematic bias of -0.09 with a proportional bias at the extreme scores. Furthermore, despite the low κw of 0.20, the overall PA at the exact and ± 0.25-unit of BCS error range between MAN_BCS and AI_BCS regarding the ΔBCS was 45.7 and 88.2%, respectively. A strong linear relationship was observed between BFT and AI_BCS (ρ = 0.75), although weaker than that between BFT and MAN_BCS (ρ = 0.91). The system was able to predict single BCS and ΔBCS with satisfactory accuracy, comparable to that obtained between trained human scorers.

Public and industry knowledge and perceptions of US swine industry castration practices

In the United States (US), surgical castration of male piglets is typically performed without any form of analgesia. This may raise concerns with the public; however, there is no information regarding current public knowledge on swine industry practices in the US. In this study we gained insight into public knowledge and perception on castration with and without analgesia in comparison to knowledge of industry stakeholders on these same topics. Through an online survey, 119 respondents were asked four questions about castration in the US swine industry. Industry respondents were contacted via social media and networking. The general public sample was accessed through Mechanical Turk. Survey responses were categorised by experience (industry vs public). Industry respondents were more aware of practices compared to the general public. Most public respondents were unaware of castration practices and the lack of analgesia use. Respondents from rural communities were more aware of castration practices than (sub)urban communities and more aware of analgesia use than those from urban communities. Those with more education had greater awareness of castration practices (occurrence not frequency). Based on the results from this first US sample, knowledge on industry practices was especially lacking for public respondents, but also for a minority of industry respondents, indicating opportunities for education and further research on the topic.

Postpartum longissimus dorsi muscle loss, but not back fat, is associated with resumption of postpartum ovarian activity in dairy cattle

The objectives of this observational cohort study were to assess the effect of body condition score change, back fat depth change, and muscle diameter change on the time to commencement of luteal activity and first estrus in commercial pedigree Holstein cows. A total of 140 of 200 commercial pedigree Holstein cows were enrolled in one dairy herd in Somerset, UK, over 52 wk in 2021 to 2022. The herd used 4 automatic milking machines with in-line progesterone measurement capability to determine commencement of luteal activity and time to first estrus. Cows were followed until at least 60 d postpartum, and milk progesterone was measured daily starting from 10 DIM. Body condition scoring and ultrasound measurements of back fat depth and longissimus dorsi muscle diameter were performed on cows twice, within 7 d of both calving and 60 DIM. Other explanatory variables assessed included parity, 60-d and 305-d milk yield, and subclinical ketosis (β-hydroxybutryate ≥1.2 mmol/L). Occurrence of clinical disease <60 DIM was forced into all models as a binary variable. Data were analyzed using multivariable Cox proportionate survival analyses. Muscle loss was associated with commencement of luteal activity and time to first estrus. A reduction in muscle diameter by 1.5 to 5 mm was associated with the shortest time to the start of luteal activity and first estrus. A reduction in muscle diameter >8 mm was associated with the longest times to luteal activity and first estrus. In addition to being affected by muscle loss, commencement of luteal activity was delayed by subclinical ketosis, clinical disease, and failure to gain body condition to 60 DIM. Cows that had a BCS loss of 0.25 or more between calving and 60 DIM were at least 52 ± 22% less likely to have commenced luteal activity compared with those that gained BCS. Interestingly, cows that had no change in body condition score commenced luteal activity later than those that gained body condition score. Muscle loss was associated with time to first estrus irrespective of clinical disease status. Cows that lost >8 mm of muscle diameter showed estrus behavior later than cows that lost 1.5 to 5 mm. In conclusion, our findings indicate that extensive muscle loss postpartum was associated with a delayed start to luteal activity and first estrus, irrespective of body condition change, clinical disease, and subclinical ketosis. Marginal muscle loss and a gain in body condition, however, were associated with an earlier start to luteal activity and first estrus.

Mean pulmonary arterial pressures in Angus steers increase from cow-calf to feedlot-finishing phases

Pulmonary arterial hypertension due to hypoxia-induced pulmonary vascular remodeling is the predominant cause of right-sided congestive heart failure (CHF) in cattle. Historically, heart failure was problematic only at altitudes over 2,134 m. However, anecdotal reports suggest that the incidence of heart failure is increasing in feedlot cattle at moderate altitude (800 to 1,600 m), with late-fed, or fat, cattle at greatest risk. The goal of this study was to evaluate pulmonary arterial pressures (PAP) in a cohort of male Angus calves from suckling to finishing to better understand why heart failure is particularly problematic in fat cattle. It was hypothesized that mean PAP would increase through the feeding period and that the calves with the greatest pressures at high altitude would have the greatest pressures as fat cattle. A total of 362 PAP measurements were obtained from 153 calves. Calves were tested at altitudes of 2,170 (4 and 6 mo old), 1,560 (13 mo old), and 1,300 m (13 and 18 mo old). Mean PAP were greater in 18-mo-old steers than any other age group (mean = 50.3 mm Hg; 95% confidence interval 48.2 to 52.4; < 0.05). Calves that had the greatest mean pressure at 6 mo of age tended to have the greatest mean pressures at 18 mo of age ( = 0.45, < 0.001). The increase in mean PAP with increasing age and adiposity likely predisposed the steers to an increased risk of CHF during the finishing phase.

Right-Sided Congestive Heart Failure in North American Feedlot Cattle

Background

Anecdotal reports suggest the incidence of right‐sided congestive heart failure (RHF) in feedlot cattle is increasing; however, the rate of occurrence and risk factors are largely unknown.

Objective

The purposes of this study were to evaluate the risk of RHF over time and among feedlots, to characterize some of the risk factors for RHF, and to investigate how risk factors may affect the timing of RHF occurrence.

Animals

The population at risk consisted of 1.56 million cattle that were placed in 10 Canadian feedlots during the years 2000, 2004, 2008, and 2012, and 5 US feedlots during the year 2012.

Methods

A retrospective observational study was conducted. Variables, including year of feedlot entry, were evaluated for association with RHF using zero‐inflated negative binomial and logistic regression models. Factors affecting time to RHF were evaluated using Cox proportional hazard regression analyzes. Death from digestive disorders (DD) served as a control.

Results

The risk of RHF in Canadian feedlots doubled from the year 2000 to the year 2012 (P = .003). For every 10,000 cattle entering US feedlots in 2012, 11 cattle died from RHF and 45 cattle died from DD. The median time to RHF was 19 weeks. Cattle treated for bovine respiratory disease were 3 times more likely to die from RHF, and they died earlier in the feeding period.

Conclusions

A doubling of the incidence of RHF over a short time period is concerning, particularly for US feedlots situated at moderate altitudes in the High Plains.

The altitude at which a calf is born and raised influences the rate at which mean pulmonary arterial pressure increases with age

Right heart failure secondary to pulmonary hypertension is a leading cause of mortality among suckling beef calves in the Rocky Mountain region. The objective of this study was to track changes in pulmonary arterial pressures (PAP) in healthy calves born and raised at altitudes ranging from 1,470 to 2,730 m. It was hypothesized that calves located at higher altitudes would show a greater increase in mean PAP (mPAP) with age than would be experienced by calves located at lower altitudes. The rationale is that high altitude hypobaric hypoxia causes a greater rate of vascular remodeling and, consequently, greater resistance to blood flow than calves located at lower altitudes. A prospective study was conducted on 5 cohorts of suckling calves from 4 herds located at altitudes of 1,470, 2,010, 2,170, and 2,730 m. In total, 470 PAP measurements were obtained from 258 calves. As hypothesized, calves located at altitudes ≥2,170 m showed a significant increase in mPAP with age ( ≤ 0.002) whereas calves at 1,470 m did not ( = 0.16). Except for calves at 2,170 m ( < 0.001), systolic PAP did not increase with age ( ≥ 0.16). Diastolic PAP increased with age at altitudes ≥ 2,170 m ( ≤ 0.09) but did not change in calves at 1,470 m ( = 0.20). In summary, mPAP and the rate at which mPAP increases with age are positively associated with the altitude at which calves are born and raised.

Technical note: A comparison of DNA collection methods in cattle and yaks

A variety of biological materials are suitable for the analysis of bovine DNA. The objective of this study was to evaluate the ease of collection, storage, and cost as well as quality and quantity of DNA samples obtained from Bos taurus (European cattle) and Bos grunniens (yak) using 2 different sample types: whole blood sampling and nasal swabs. Hair follicle DNA samples from yaks were also analyzed. Deoxyribonucleic acid samples were collected from 1 herd of Black Angus yearling bulls (n = 166) and 1 herd of yaks (n = 24). A NanoDrop Bioanalyzer ND1000 was used to quantify DNA. To assess DNA purity, absorbance ratios were determined at wavelengths of 260 nm relative to 280 nm and 260 nm relative to 230 nm. Single nucleotide polymorphism genotyping was performed using a competitive allele-specific PCR (KASP) genotyping system and the call rates to 3 specific SNP were compared. Using a commercially available nonautomated ethanol DNA extraction technique, nasal swabs yielded a greater quantity of DNA than blood (P < 0.0001) and a greater quality DNA sample than blood (P < 0.0001). Blood and nasal swab performance in SNP genotyping assays were similar (P = 0.5). The greater expense of nasal swabs was offset by their ease of use: less time, skill, and equipment was needed to obtain a sample and the storage of samples was more convenient (room temperature). In yaks, accessing the coccygeal vein, which is relatively straightforward in cattle, was difficult. Nasal swabbing and hair follicle sampling in yaks was performed relatively easily. Yak hair follicles were a poor source of DNA. In conclusion, DNA collection using nasal swabs was more convenient and provided a greater quantity of DNA and better quality sample than blood collection in both Angus and yak. Notably, yak hair was a poor source of DNA, and yak blood was difficult to obtain.

Antibodies to loop 6 of the P2 porin protein of nontypeable Haemophilus influenzae are bactericidal against multiple strains

The P2 porin protein is the most abundant protein in the outer membrane of nontypeable Haemophilus influenzae (NTHI). Analysis of sequences of P2 from different strains reveals the presence of both heterogeneous and conserved surface-exposed loops of the P2 molecule among strains. The present study was undertaken to test the hypothesis that antibodies to a conserved surface-exposed loop are bactericidal for multiple strains of NTHI and could thus form the basis of vaccines to prevent infection due to NTHI. Polyclonal antiserum to a peptide corresponding to loop 6 was raised and was immunopurified over a loop 6 peptide column. Analysis of the antibodies to whole organisms and peptides corresponding to each of the eight loops of P2 by immunoassays revealed that the antibodies were highly specific for loop 6 of P2. The immunopurified antibodies bound to P2 of 14 of 15 strains in immunoblot assays. These antibodies to loop 6 demonstrated complement-mediated bactericidal killing of 8 of 15 strains. These results support the concept of using conserved regions of the P2 protein as a vaccine antigen.

Hypercapnic ventilation response in patients with lung disease: improved accuracy by correcting for ventilation ability

The hypercapnic ventilation response (HCVR) is positively correlated with forced expired volume in 1 s (FEV1). Therefore, subjects of small stature or patients with lung disease have low values for HCVR. However, indexing the HCVR for the subject's predicted maximal voluntary ventilation (MVV) results in a corrected HCVR (CHCVR) which is not dependent on FEV1 in normal subjects [Respiration 1993;60:197-202]. We hypothesized that the CHCVR would also be useful in assessing chemosensitivity in patients with poor lung function. To obtain the predicted MVV, we used the linear regression for FEV1 vs. measured MVV obtained from 411 patients with a wide range of FEV1 values (MVV = 31.2 x FEV1 + 11.8, r = 0.90, p < 0.001). We compared HCVR and CHCVR to the occlusion pressure response to hypercapnia (OPRH) in 34 patients with chronic obstructive pulmonary disease (COPD) and in 19 patients with low FEV1 due to small stature. All patients had been referred for assessment of possible sleep apnea. The results for the two groups of patients were similar. For the COPD patients, the HCVR had high values for sensitivity (86%) and negative predictive value (94%), but specificity, positive predictive value and accuracy were low (59, 35 and 65%, respectively). In contrast, CHCVR had high values for all the foregoing (86, 96, 100, 100 and 97%, respectively). Our results suggest that the CHCVR is useful in assessing chemosensitivity in patients who are ventilation-limited.

Normal values for the hypercapnic ventilation response: effects of age and the ability to ventilate

Normal values reported for the hypercapnic ventilation response (HCVR) vary considerably, but the reported normal values have come from studies containing small sample sizes and/or the subjects were young or of unidentified age. We speculated that age has a major effect on HCVR due to the generally lower maximal ventilation (VEmax) of elderly subjects. Therefore, we performed a large study to more accurately define the normal range and to reveal any effects which age and sex might have on HCVR. We studied 181 normal subjects (69 males, 112 females) between the ages of 20 and 93 years. Prior to measuring HCVR we measured forced expired volume in 1 s (FEV1) and forced vital capacity to establish whether lung function was normal and to obtain an estimate of VEmax. Results for the entire group revealed a significant correlation between FEV1 and HCVR [HCVR = 0.51 + (0.33 FEV1), r = 0.43, p < 0.001], so it is clear that the ability to ventilate can influence HCVR. We also found a significant correlation between age and HCVR [HCVR = 2.08-(0.01.age), r = 0.34, p < 0.001]. The mean HCVR for our male group (1.86 +/- 0.54 l/min.mm Hg) was significantly higher (p < 0.001) than that for the females (1.37 +/- 0.60) even though their mean ages were similar (41.3 vs. 41.7 years, respectively). We expect that this difference in HCVR was due to the higher FEV1 in males compared to females (3.82 vs. 2.83 liter, respectively).(ABSTRACT TRUNCATED AT 250 WORDS)