Preprint: genomic SEM identifies shared and divergent genetic architecture for locomotor activity across rat studies

A preprint posted to bioRxiv reports a large-scale genetic analysis of locomotor activity in heterogeneous stock (HS) rats — a multiparental outbred population used in addiction and behavioural genetics research. The study draws on 9,341 male and female rats from eight independent studies, exploiting the large combined sample to characterise the genetic architecture of novelty-induced locomotor response, a heritable trait associated with addiction vulnerability in rodents.

The authors found that locomotor activity was heritable within studies, and that pairwise genetic correlations across studies were consistently positive but varied in magnitude — indicating that the studies are measuring related but not identical phenotypes, potentially reflecting differences in environmental context, sex ratios, or measurement protocol.

To address this non-uniform architecture, the team compared three complementary locus-discovery strategies: a pooled analysis across studies (MEGA), a fixed-effects meta-analysis of per-study GWAS results, and genomic structural equation modelling (Genomic SEM), which can model a latent common factor underlying the correlated phenotypes while allowing study-specific residual variation. Genomic SEM identified additional loci not detected by the other approaches, and the authors argue this approach is better suited to multi-study behavioural data with heterogeneous genetic architectures.

The work is relevant to researchers in statistical genetics, quantitative genetics, and the neurobiology of addiction. It has not yet been peer-reviewed.