Environmental Association Analysis
Temperature and precipitation are considered the driving factors that
affect the population growth rate and limit the distribution of species
(Cahill et al., 2014; Dalgleish, Koons, Hooten, Moffet, & Adler, 2011;
Kim & Donohue, 2013). Therefore, nineteen current bioclimatic variables
with a spatial resolution of 30 s were collected from WorldClim
(http://www .worldclim.org/).Simultaneously,
we recorded the GPS information of the sampling locations and downloaded
the GPS information of A. viridiflora from CVH
(http://www.cvh.ac.cn) and GBIF
(http://www.gbif.org). ArcMap v.10.4 was used
to limit the spatial extent according to the buffer radius (5 km) around
each occurrence record. We used |r| < 0.8
(Pearson correlation coefficient) as a cutoff to remove highly
correlated variables. The seven retained current bioclimatic factors
(Bio1: annual mean temperature; Bio2: mean diurnal range; Bio3:
isothermality; Bio4: temperature seasonality; Bio8: mean temperature of
the wettest quarter; Bio15: precipitation seasonality; Bio17:
precipitation of the driest quarter) were used for subsequent analysis.
Redundancy analysis (RDA) was used to assess the impact of current
bioclimatic factors on the genomic composition of the A.
viridiflora complex in R. For the same reason as selection analysis, we
also used CN and NW lineages to identify the loci related to the seven
retained current bioclimatic factors. BAYENV2 (Coop, Witonsky, Di
Rienzo, & Pritchard, 2010) was used with 1,000,000 iterations and run
three times separately.
For
each bioclimatic factor, the SNPs among the top 1% according to BF and
among the top 5% according to the absolute Spearman’s ρ were considered
candidates. Candidate SNPs were mapped to the corresponding genes, and
GO enrichment was performed by the clusterProfiler package (Yu et al.,
2012) in R.