Experimental support for genomic prediction of climate maladaptation using the machine learning approach Gradient Forests.

摘要: Gradient Forests (GF) is a machine learning algorithm that gaining in popularity for studying the environmental drivers of genomic variation and incorporating information into climate change impact assessments. Here we (i) provide first experimental evaluation ability "genomic offsets" - metric maladaptation derived from to predict organismal responses change, (ii) explore use GF identifying candidate SNPs. We used high-throughput sequencing, genome scans, several methods, including GF, identify loci associated with adaptation balsam poplar (Populus balsamifera L.). Individuals collected throughout poplar's range also were planted two common garden experiments. relate gradients expected magnitude response (i.e., genetic offset maladaptation) populations when transplanted their "home" environment gardens. then compared predicted offsets different sets randomly selected SNPs measurements population performance found inverse relationship between performance: larger performed worse gardens than smaller offsets. Also, better did "naive" transfer distances. However, slightly Our study provides evidence represent order estimate degree exposed rapid suggests may have some promise as method