Gene annotation for the significantly associated SNPs
Of the 1458 SNPs associated with environmental gradients, functions could be assigned for 788 (54%), while the rest had no matches in available gene ontology databases. We found that 283 SNPs with identifiable functions belonged to protein types that may be directly related to drought tolerance or other environmental responses (Fig. 4). We categorized these genes into five main functional groups: (a) the ubiquitination pathway, (b) seed, pollen and ovule formation, (c) cell wall formation, (d) stress responses, and (e) cell division and growth. Other associated SNPs with known functions were in or near transcription factors and genes with expression-regulating functions.
Many of the SNPs associated with TMAX, TMIN, CWD, and PCK4 were in or near genes in the protein ubiquitination pathway or the jasmonic acid synthesis response pathways (Fig. 4 and Table S2), both of which are involved in responses to biotic or abiotic stress (Creelman & Mullet 1995; Lyzenga & Stone 2012; Stone 2014). CWD and PCK4 were also associated with SNPs in or near genes involved in seed dormancy, cell wall organization, and the abscisic acid (ABA) signaling pathway, which have been previously linked to drought responses in trees (Moranet al. 2017b). Genes involved in reproduction, including pollen and ovule formation, were associated with TMAX, TMIN, and PCK4. Genes involved in vascular tissue formation, growth regulation, and stress responses were associated with TMAX and PCK4. Genes involved in stomatal regulation and pathogen responses were associated with TMIN and PCK4. Further biotic and abiotic stress response genes were associated with PCK4, as were genes involved in nutrient transport, photosynthesis, respiration, sugar synthesis, and light responses (Table S2).
Of the 817 SNPs associated with seedling control (wet treatment) trait values and 1,154 SNPs associated with trait drought responsiveness, 43% and 51% could be assigned functions by gene ontology (Additional file 2: Table S3 and Table S4). Many of the same functional categories of genes associated with the environment were also related to measured phenotypes. This includes ubiquitination, seed development, cell wall organization, stress response, cell division (Fig. 4, 5, 6), and transcription factors. However, there was no overlap in specific SNPs identified.
The control treatment levels of the two stomatal traits were associated with genes involved in ubiquitination, cell wall organization or modification, growth and development, and ABA response. Control root-to-shoot ratio was associated with genes involved in biotic & abiotic stress responses, cell wall organization or modification, cell division or differentiation, lateral root formation, and ubiquitination. Control height growth had no associated SNPs, and root length was only associated with one SNP located in a gene involved in ubiquitination (Fig. 5). However, drought responsiveness of height growth, shoot weight, and root length were associated with all five functional categories (Fig. 6). Drought responsiveness of the two stomatal traits was associated with genes involved in stress responses, cell wall formation/organization, cell division/differentiation, and root formation.
Besides the five main functional groups of genes with SNPs associated with climatic, phenotypic, and drought response variables, several other functional groups were identified in the GEA and GPA annotation results (Table S2, S3, and S4). For example, 111 (14%) of the environmentally associated SNPs, 53 (6%) of SNPs associated with control traits, and 121 (12%) of the SNPs associated with trait drought responses were in genes relating to ATP binding or protein kinases. It was also fairly common for associated SNPs to be in genes associated with RNA/DNA binding, metal ion binding, translation, and protein transport.