Meta-analysis for selection of feed efficiency in crossbred pigs

We reviewed the literature and completed a meta-analysis of parameter estimates to determine which current and novel traits are promising indicator traits to select for feed efficiency of crossbred pigs. Feed conversion ratio (FCR) was the key trait of interest, average daily gain and daily feed intake were also included. For each following category a single trait was selected; digestibility (dry matter), microbiota (Alloprevotella), feeding behaviour (eating time per day), group records (group daily feed intake), welfare (joint lesions or total lesion count), indirect genetic effects (growth rate with social effect), and biomarkers (nitrogen excreted). A genetic correlation matrix was built with published results, and unknown correlations assumed as low (0.10). The nearest positive definite matrix and published genetic standard deviations were used to calculate a covariance matrix. To determine which traits have the most potential to improve FCR, 100% of selection was placed on each trait separately, until the selected trait increased by one genetic standard deviation. When selection was placed on FCR, it decreased by 7.1%, from 2.52 (kg/kg) to 2.34 (kg/kg). The other traits reduced FCR in descending order; dry matter digestibility (4.5%), daily feed intake (2.2%), average daily gain (1.9%), eating time per day (1.3%), nitrogen excreted (1.2%), group daily feed intake (0.8%), total lesion count (0.8%), Alloprevotella (0.7%), joint lesions (0.7%), and growth rate with social effect (0.7%). Current traits are likely to have the largest impact on feed conversion ratio. Including novel traits in selection, is likely to have a positive impact on FCR, particularly from; digestibility, feeding behaviour, biomarkers, and group records. To determine an optimal index to select for feed efficiency, further analysis will use selection index theory with multi-trait selection. It will include additional traits from each category (including perturbations as a new category), estimate missing correlations with industry data, account for differences in genetic correlations between purebreds and crossbreds, and take into consideration ease and costs of measuring to prioritise traits.