Statistical and machine learning approaches to describe factors affecting preweaning mortality of piglets

Authors: RAHMAN, MD TOWFIQUR - University Of Nebraska; BROWN-BRANDL, TAMI - University Of Nebraska; Rohrer, Gary; SHARMA, S. RAJ - University Of Nebraska; MANTHENA, VAMSI - University Of Nebraska; SHI, YEYIN - University Of Nebraska

Submitted to: Translational Animal Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/3/2023
Publication Date: 10/25/2023
Citation: Rahman, M., Brown-Brandl, T.M., Rohrer, G.A., Sharma, S., Manthena, V., Shi, Y. 2023. Statistical and machine learning approaches to describe factors affecting preweaning mortality of piglets. Translational Animal Science. 7(1). Article txad117. https://doi.org/10.1093/tas/txad117.
DOI: https://doi.org/10.1093/tas/txad117

Interpretive Summary: High pre-weaning mortality (PWM) rates for piglets are a significant concern for the worldwide pork industries. This issue is an economic loss for producers and a well-being concern for the industry. Factors contributing to PWM include number of mummies and stillborns, health condition of sows and piglets, environmental factors, congenital abnormalities, and overlaying by sow. Understanding which factors contribute the most to PWM will enable producers and scientists to focus their efforts in specific areas to improve production efficiency. The present study focuses on determining the effects of different factors on the occurrence of PWM and predicting PWM using machine learning models. Data were collected from 1,982 litters located at the US Meat Animal Research Center (USMARC). Production records for each litter included parity, season, location of the pen, gestation length, litter line (Yorkshire, Landrace sired), health records, number of piglets stillborn, piglets born alive, litter size, and mean birth weight. On average, the mean birth weight was 1.44 kg, mortality was 16.1% and overlay percentage was 6.2%. There were no significant effects found for season and location in a room on PWM. The effects for litter lines were significantly different for PWM and percent overlays. Landrace-sired piglets have 17.1% PWM and 6.5% overlays which is greater than Yorkshire-sired PWM of 15.7% and overlay of 6.4%. PWM increased with greater parity with fourth parity sows having PWM of 18.2%. Low birth weight and greater litter size significantly increased mortality. Most important factors to predict PWM were litter size, mean birth weight, number of health diagnoses, gestation length and parity. Considering the challenge to reduce pre-weaning mortality and limited studies exploring multiple major contributing factors, this study is valuable for creating a greater understanding of the inter-related factors contributing to PWM in commercial swine production. Increasing mean birth weight and gestation length while maintaining high health standards are important areas of research that should be conducted to reduce piglet mortality in commercial swine production. In addition, greater oversight of older parity sows from birth until 3 days post-farrowing could improve survival of young piglets.

Technical Abstract: High preweaning mortality (PWM) rates for piglets are a significant concern for the worldwide pork industries, causing economic loss and well-being issues. This study focused on identifying the factors affecting PWM, overlays, and predicting PWM using historical production data with statistical and machine learning models. Data were collected from 1,982 litters from the U.S. Meat Animal Research Center, Nebraska, over the years 2016 to 2021. Sows were housed in a farrowing building with three rooms, each with 20 farrowing crates, and taken care of by well-trained animal caretakers. A generalized linear model was used to analyze the various sow, litter, environment, and piglet parameters on PWM. Then, different models (beta-regression and machine learning model: a random forest [RF]) were evaluated. Finally, the RF model was used to predict PWM and overlays for all listed contributing factors. On average, the mean birth weight was 1.44 kg, and the mean mortality was 16.1% where 5.55% was for stillbirths and 6.20% was contributed by overlays. No significant effect was found for seasonal and location variations on PWM. Significant differences were observed in the effects of litter lines on PWM (P < 0.05). Landrace-sired litters had a PWM of 16.26% (±0.13), whereas Yorkshire-sired litters had 15.91% (±0.13). PWM increased with higher parity orders (P < 0.05) due to larger litter sizes. The RF model provided the best fit for PWM prediction with a root mean squared errors of 2.28 and a correlation coefficient (r) of 0.89 between observed and predicted values. Features’ importance from the RF model indicated that, PWM increased with the increase of litter size (mean decrease accuracy (MDA) = 93.17), decrease in mean birth weight (MDA = 22.72), increase in health diagnosis (MDA = 15.34), longer gestation length (MDA = 11.77), and at older parity (MDA = 10.86). However, in this study, the location of the farrowing crate, seasonal differences, and litter line turned out to be the least important predictors for PWM. For overlays, parity order was the highest importance predictor (MDA = 7.68) followed by litter size and mean birth weight. Considering the challenges to reducing the PWM in the larger litters produced in modern swine industry and the limited studies exploring multiple major contributing factors, this study provides valuable insights for breeding and production management, as well as further investigations on postural transitions and behavior analysis of sows during the lactation period.