Longitudinal genomic analyses of automatically-recorded vaginal temperature in lactating sows under heat stress conditions based on random regression models

Abstract

Background: Automatic and continuous recording of vaginal temperature (T_V) using wearable sensors causes minimal disruptions to animal behavior and can generate data that enable the evaluation of temporal body temperature variation under heat stress (HS) conditions. However, the genetic basis of T_V in lactating sows from a longitudinal perspective is still unknown. The objectives of this study were to define statistical models and estimate genetic parameters for T_V in lactating sows using random regression models, and identify genomic regions and candidate genes associated with HS indicators derived from automatically-recorded T_V.

Results: Heritability estimates for T_V ranged from 0.14 to 0.20 over time (throughout the day and measurement period) and from 0.09 to 0.18 along environmental gradients (EG, - 3.5 to 2.2, which correspond to dew point values from 14.87 to 28.19 ˚C). Repeatability estimates of T_V over time and along EG ranged from 0.57 to 0.66 and from 0.54 to 0.77, respectively. T_V measured from 12h00 to 16h00 had moderately high estimates of heritability (0.20) and repeatability (0.64), indicating that this period might be the most suitable for recording T_V for genetic selection purposes. Significant genotype-by-environment interactions (GxE) were observed and the moderately high estimates of genetic correlations between pairs of extreme EG indicate potential re-ranking of selection candidates across EG. Two important genomic regions on chromosomes 10 (59.370-59.998 Mb) and16 (21.548-21.966 Mb) were identified. These regions harbor the genes CDC123, CAMK1d, SEC61A2, and NUDT5 that are associated with immunity, protein transport, and energy metabolism. Across the four time-periods, respectively 12, 13, 16, and 10 associated genomic regions across 14 chromosomes were identified for T_V. For the three EG classes, respectively 18, 15, and 14 associated genomic windows were identified for T_V, respectively. Each time-period and EG class had uniquely enriched genes with identified specific biological functions, including regulation of the nervous system, metabolism and hormone production.

Conclusions: T_V is a heritable trait with substantial additive genetic variation and represents a promising indicator trait to select pigs for improved heat tolerance. Moderate GxE for T_V exist, indicating potential re-ranking of selection candidates across EG. T_V is a highly polygenic trait regulated by a complex interplay of physiological, cellular and behavioral mechanisms.

Fig. 1

Number of records and average vaginal temperature throughout the day and along environmental gradients. a Number of records (green bars) and average vaginal temperature (T_V) (red dots) for time points (every 10 min) throughout the day and b number of records (green bars) and average T_V (red dots) along environmental gradients based on dew point values

Fig. 2

(a left column) Additive genetic variance, permanent environmental variance, residual variance, heritability, and repeatability along the time trajectory throughout the day and (b right column) continuous environmental gradient for vaginal temperature of lactating sows based on random regression model analyses. Red dot points (●) represent models considering heterogeneous residual variance and green triangles (▲) represent models considering homogeneous residual variance

Fig. 3

Estimates of genetic correlations (Cor) for vaginal temperature in lactating sows under heat stress between a pairs of time points (each 10 min) throughout the day (range: 0.76–1) and b pairs of environmental gradients (standardized dew point) (range: 0.44–1)

Fig. 4

Venn diagram of the count of candidate genes identified based on a heat stress (HS) time-periods and b environmental gradient class. The four potential HS time-periods throughout the day were defined as (1) from 23h00 to 06h30, representing when vaginal temperature (T_V) starts to decrease with decreased ambient temperature (T_a) and with relatively more comfortable environmental conditions; (2) from 06h30 to 09h30, representing T_V maintained at a relatively low level when environment is not too hot for the sows; (3) from 09h30 to 18h30, representing the time in which T_V starts to increase with increasing T_a; (4) from 18h30 to 23h00, representing T_V maintained at a relatively high level when T_a is usually above thermoneutral conditions. Each two standardized Dew point units was considered as an environmental gradient class (no to mild HS: − 3.5 to 1.5; moderate HS: − 1.5 to 0.5; severe HS: 0.5–2.5)