Mon, Aug 15, 2022: 5:00 PM-6:30 PM
ESA Exhibit Hall
Background/Question/MethodsMany species of pine, including the lodgepole pine (Pinus contorta var. latifolia) have been infested by the wood boring species, mountain pine beetle (MPB: Dendroctonus ponderosae). Anthropogenic factors have expanded the range of MPB in Canada, resulting in host-expansion to jack pine (Pinus banksiana) through a lodgepole x jack pine hybrid zone in Alberta. Jack pine is a common tree species throughout the Canadian Boreal Forest, and its range overlaps with other potential host species (e.g., red pine; Pinus resinosa), thus the threat of continued expansion may have extensive ecosystem effects across Canada. Knowledge regarding the epidemic potential of MPB in jack pine, where host choice of MPB may result in limited attack success in jack pine is limited. This study aimed to model the geographic spread of MPB based on genetic ancestry of the host species. Using highly accurate predictions of pine ancestry across the landscape, we estimated the ancestry of individuals genotyped at 28 species discriminating genetic loci to use as the response variable. We compared two different models, a logistic regression, and an ordinal categorical approach, using crucial host specific environmental variables as predictors. A second genetic dataset was used to compare the two models for accuracy.
Results/ConclusionsFor the model development, we included 862 individuals, and found that different predictors were important for the two different models, but both models included variables associated with temperature, moisture and location. Using an additional 1,242 samples, our preliminary results suggest that both models are highly accurate (AUC ~ 0.94), however, the ordinal regression model performs better for the hybrid category. Mapping the prediction of these two models allowed us to compare areas of disagreement/agreement. Our next steps will be to ground truth these models using genotyping data from Alberta seedlots collected within areas of transition between the species classes. This research will improve our understanding of the lodgepole x jack pine hybrid zone and will help fill a key knowledge gap that is necessary in order to access the eastern spread risk of MPB.
Results/ConclusionsFor the model development, we included 862 individuals, and found that different predictors were important for the two different models, but both models included variables associated with temperature, moisture and location. Using an additional 1,242 samples, our preliminary results suggest that both models are highly accurate (AUC ~ 0.94), however, the ordinal regression model performs better for the hybrid category. Mapping the prediction of these two models allowed us to compare areas of disagreement/agreement. Our next steps will be to ground truth these models using genotyping data from Alberta seedlots collected within areas of transition between the species classes. This research will improve our understanding of the lodgepole x jack pine hybrid zone and will help fill a key knowledge gap that is necessary in order to access the eastern spread risk of MPB.