Past stress responses archived in tree-rings associate with SNP genotypes in Abies alba (Mill.)

Air pollution, especially containing sulfur dioxide (SO2), is suspected to be the main contributor to the foliar damage and dieback of silver fir ( Abies alba Mill.) populations in the 1970s and 1980s in Germany. In combination with an increased sensitivity to drought, caused by SO 2 , this led to a marked decrease in radial growth in many silver fir trees. This growth depression pe- riod is archived in the annual tree-ring data, which is usually studied on the population level. We derived 'dendrophenotypes' that characterize resistance, resilience and recovery during the depres- sion period based on individual tree-ring widths of silver fir trees from stands at two elevations in the Bavarian Forest National Park in Germany. Our goal was to associate genetic variation, in the form of SNPs, with variation in the dendrophenotypes to identify candidate genes for potentially adaptively relevant stress responsive traits in silver fir. Using feature selection techniques based on the machine learning algorithm random forest, we could identify 15 candidate genes whose products are mostly involved in photosynthetic or chloroplast development and some in drought response. This shows that individual-level dendrophenotypes are a valuable measure for genetic association studies in forest trees and can strongly increase our understanding of the genetic basis of environmental stress response, specifically to extreme episodic events