| The Tibetan Plateau is the highest plateau in the world,its unique ecological environment shapes local plant characters and evolution profoundly.Previous studies have shown that high and low altitudes have different degrees of selection on different sites in the plant genome,which results in the diversity of plant growth patterns.To better understand the genetic basis of plants’ adaptability to high altitudes,most studies focus on means such as genome sequencing,comparative genomics and transcriptome sequencing,whereas few studies were from the perspective of genetic regulatory network(GRN).Populus szechuanica var.tibetica is a native tree species distributed on the Tibetan Plateau.For a long time,its adaptation to the plateau environment has made it an ideal material for research on environmental adaptability and stress resistance traits.In order to fully analyze the genetic mechanism of plants adaption to different altitudes,this study collects samples of Populus szechuanica var.tibetica from high and low altitude to build two natural populations respectively.According to the experimental design of the homogeneous garden,the growth traits of all individuals of each population are measured annually generalized.An inferring informative,dynamic,omnidirectional,and personalized Fun Graph is introduced to construct growth trait related GRNs of high and low altitude populations respectively,which are to reveal the mechanism of gene interaction regulation on growth variation,and to further interpret the response mechanism of growth varition under environmental stress.The main results and conclusion are listed as follows:(1)Using the allometric growth equation to fit the ground diameter growth and plant height growth of Populus szechuanica var.tibetica at different altitudes and regions,the goodness of fit was greater than0.7,and the fitting effect was good.(2)The ground diameter growth and plant height growth of Populus szechuanica var.tibetica at different altitudes and different regions were analyzed by functional mapping,and the threshold was determined by Bonferroni test.At the high altitude in Dayi,167 SNPs were mapped to QTLs related to ground diameter growth,and 34 genes were annotated;440 SNPs were related to plant height growth QTLs,and 85 genes were annotated.At the low altitude in Dayi,8 SNPs were mapped to QTL related to ground diameter growth,and 1 gene was annotated;741 SNPs were related to plant height growth,and123 genes were annotated.At the high altitude in Nantong,215 SNPs were mapped to QTLs related to ground diameter growth,and 41 genes were annotated;501 SNPs were related to plant height growth,and96 genes were annotated.At the low altitude in Nantong,3 SNPs were mapped to QTL related to ground diameter growth,and 1 gene was annotated;369 SNPs were related to plant height growth,and 75 genes were annotated.(3)Fun Graph was used for the first time to reconstruct a genome-wide multilayer interaction network of P.According to the two phenotypes at different altitudes in different regions,a total of 8 groups of Fun Graph gene-wide multi-level networks were established,the genetic effects were calculated,and clustered by Fun Clu,the cluster with the smallest BIC value was selected from 2 to 400 categories according to the BIC standard.For optimal clustering,a network is established between modules in the first layer.For example,in the growth difference network of Populus szechuanica var.tibetica in the highaltitude area of Nantong area,the best clustering was 176 categories determined by BIC,and the number of network connections accounted for 1.67% of the theoretical number of connections.The network has good sparsity,contains 44 QTL modules,and the functions of each module are enriched and analyzed.In the second layer network,repeat the clustering and BIC screening process to build a submodule network(M33)or build a gene network between SNP-SNPs(M113),and then build a third layer network(SM3,etc.),until it cannot continue.so far.In sub-submodule SM3 of subnetwork module M33,a detailed map of the regulatory network associated with vacuolar processing enzymes is provided,illustrating how this QTL works with all other QTL/SNPs to determine growth differences through genetic interactions.This clustering enables us to reconstruct multiple interconnected small networks from a large network to overcome the curse of network dimensionality.The SNP-SNP network of growth variation can be annotated by coding genes for protein roles.With the QTL regulatory mechanisms revealed in the model,our multilayered genetic network provides a systematic tool for dissecting the evolution of homogeneous environments at different altitudes. |
