In Porcine health management
BACKGROUND : Knowing the feed intake pattern during lactation of modern genetic sows is crucial because it allows to anticipate possible problems and maximize their performance. On the other side, electronic feeders permit real-time data to be available for a more accurate evaluation of sow eating behavior. This work aimed to characterize the feed intake patterns of lactating highly prolific sows and determine their effect on reproductive performance. A database of 1,058 registers of feed intake collected from a commercial farm was used to identify five consistent sets of clusters (feeding curves) using machine learning. In the second step, the five feeding curves were characterized into five patterns by high, medium and low feed intake during 0-6 d and 7-28 d of lactation: 1-HH, 2-MH, 3-HM, 4-MM and 5-LL.
RESULTS : The mean daily feed intake of all the sows was 6.2 kg (0.06 SEM) across the 5 patterns. As the pattern numbers increased from 1-HH, 2-MH, 3-HM and 4-MM to 5-LL, their mean daily feed intake decreased from 7.6 to 6.9, 6.4, 5.8 and 4.3 (0.06 SEM) kg, respectively (P < 0.01). Sows with Pattern 1-HH tended to have shorter weaning-to-first service interval (P = 0.06) and had a higher farrowing rate than those with Pattern 5-LL (P < 0.01). Furthermore, contrast analysis showed that sows with Patterns 1-HH and 2-MH tended to have more piglets weaned (P = 0.05) and lower preweaning mortality (P = 0.07) than those with Patterns 3-HM and 4-MM. Also, sows with Patterns 1-HH and 3-HM had fewer stillborn piglets and a lower percentage of stillborn piglets and mummies than those with Patterns 2-MH and 4-MM (P < 0.01).
CONCLUSIONS : This study indicates the importance of reaching Pattern 1-HH by rapidly increasing feed intake during early lactation and high feed intake during late lactation, which is associated with high weaning performance and subsequent reproductive performance of the sows. Also, the current study suggests that Pattern 1-HH is linked to good farrowing with a low percentage of stillborn piglets and mummies. Finally, it is critical for producers to timely identify a problem of sows' eating behavior and to make a prompt decision to intervene.
Rodríguez María, Díaz-Amor Gonzalo, Morales Joaquín, Koketsu Yuzo, Piñeiro Carlos
2023-Jan-31
Commercial swine herds, Feed consumption, Feeding behavior, Lactation, Machine learning, Reproductive performance
