Data reconciliation in multi-trait experiments with kinship ordination

Authors: VAN EE, JUSTIN - Colorado State University; VAHSEN, MEGAN - University Of Georgia; GAMBA, DIANA - Pennsylvania State University; MAXWELL, TOBY - Oregon State University; DAVIDSON, BILL - Us Geological Survey (USGS); LAZARUS, BRYNNE - Us Geological Survey (USGS); Porensky, Lauren; HUFBAUER, RUTH - Colorado State University; GERMINO, MATTHEW - Us Geological Survey (USGS); ADLER, PETER - Utah State University; LASKY, JESSE - Pennsylvania State University; HOOTEN, MEVIN - University Of Texas At Austin

Submitted to: Methods in Ecology and Evolution
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/16/2026
Publication Date: 4/14/2026
Citation: Van Ee, J.J., Vahsen, M.L., Gamba, D., Maxwell, T.M., Davidson, B.E., Lazarus, B.E., Porensky, L.M., Hufbauer, R., Germino, M.J., Adler, P.B., Lasky, J.R., Hooten, M.B. 2026. Data reconciliation in multi-trait experiments with kinship ordination. Methods in Ecology and Evolution. https://doi.org/10.1111/2041-210x.70284.
DOI: https://doi.org/10.1111/2041-210x.70284

Interpretive Summary: Motivated by common garden, growth chamber, and physiological experiments of invasive cheatgrass (Bromus tectorum), we extend the methods of probabilistic principal components analysis to incorporate genetic marker data and plant trait information collected across many environments. We clarify the connection between our approach and more traditional statistical approaches for multi-trait data. We demonstrate through analyses of real and simulated data that our approach generally outperforms the standard mixed model approach. Our approach provides biologically valuable information related to trait correlations and selection as simple model outputs. Summarizing this output, we gain insights into the strength of several environmental and genetic trends driving cheatgrass invasion dynamics. Adaptation of this framework could prove useful for predicting the future spread of cheatgrass in the Intermountain West of the United States.

Technical Abstract: 1. A central aim in biology is understanding the heritability of traits and how trait interactions contribute to success in diverse environments. Experiments that record multiple traits of related individuals in distinct environments are key to addressing this aim. Mixed modeling approaches have been proposed for analyzing multivariate trait data. The parameter space of these mixed models grows exponentially with the number of traits and environments considered, which increases computational demand and the risk of overfitting. Existing approaches can also be challenging to implement for datasets in which different traits were measured in different environments. 2. We developed a probabilistic principal components model that incorporates genetic marker data for estimating heritability and improving predictions of traits. Our approach promotes model parsimony by estimating covariate associations and genetic variances for a reduced number of trait principal components. Our approach accommodates variation in trait correlations across environments and can be applied in settings where only a subset of traits are observed in each environment, maximizing use of the data. 3. In a simulation study, we showed that our approach reduces bias and variability in heritability estimates and improves predictive performance relative to standard multivariate mixed modeling approaches. We applied our approach to reconcile partially overlapping datasets collected from growth chamber and common garden experiments of Bromus tectorum, an annual grass invasive to the United States. Fitting mixed models independently to the data sources resulted in biologically unreasonable estimates of narrow sense heritability, whereas a joint analysis with our probabilistic principal component model provided more reasonable inference and improved predictive performance. Our approach provides trait ordinations and selection coefficients as model derived quantities. Collecting inferences from these derived quantities, we present a holistic explanation for the strength of several clines in Bromus tectorum and discuss the relevance of these clines for invasion in the Intermountan West. 4. The flexibility, tractability, and performance of our approach make it appealing for joint inference and prediction in experiments of multiple traits collected across several environments.