Fig. 6. Effective migration patterns of Opisopappus taihangensis and Opisopappus longilobus .
a: Posterior mean migration rates m (on the log10 scale); b: Posterior mean diversity rates q (on the log10 scale)

3.3 The relationship between SNPs and landscape factors

A total of 29 SNPs were identified under selection based on theF ST outlier approach implemented in BayeScan software (q-value < 0.01) (Support files 3).
Samßada results showed that, in the first 29 (AIC < 17) effective models, seven landscape factors of Taihang Mountains, average precipitation in August, average precipitation in November, built-up land (residential and infrastructure), rain-fed cultivated land, workability (restricted site management), solar radiation in August, and soil PH, were closely related with the selected SNPs, which accounted for 13.8% of total variation revealed by BAYESCAN. Average precipitation in October in Taihang Mountains accounted for 3.4%, was significantly correlated with 29 SNP loci (Support files 5).
LFMM analysis revealed that 27 of 29 SNP loci were associated with average precipitation in August (8.66%), average precipitation in November (15.75%), built-up land (residential and infrastructure) (11.81%), rain-fed cultivated land (11.81%), workability (restricted site management) (11.81%), solar radiation in August (12.6%), soil PH (14.17%), average precipitation in October (13.39%) (Table 3).
In total, 29 selected SNPs were strongly associated with landscape features of Taihang Mountains revealed by Samβada and LFMM analysis, especially average precipitation in November.
The identified selected SNPs were mainly enriched in Carbohydrate metabolism, Energy metabolism, Translation, Signal transduction and Transport and catabolism. The genes related with these SNPs regulate Glycolysis / Gluconeogenesis, Pentose and glucuronate interconversions, Ribosome, MAPK signaling pathway. In particular, MAPK signaling and plant hormone signal transduction pathway were found to be involved in the three genes based on samβada analysis.

4. Discussion

4.1 Landscape features affect genetic characteristics of Opisthopappus species

The genetic diversity of O. longilobus was higher than that ofO. taihangensis , which is consistent with previous studies (Ye et al., 2021). Being an ancestral species of Opisthopappus , more genetic variation might be acculmated into this species than its descendant O. taihangensis . Furthmore, the OLb group of O. longilobus had the highest genetic diversity than other three groups. It suggested that this group might be a diversity centre for O. longilobus, even for the whole Opisthopappus genus, especially the areas of HLT and THDXG populations, which had the highest genetic diversity among all populations (Ren et al., 2022) (Table 1).
For the two groups of O. longilobus , the OLa group had a relative lower genetic diversity than OLb group (Table 1). The populations of OLa group are at the margins of geographic range. Generally, the marginal populations are often thought to be poorly adapted to their environment (Bontrager and Angert, 2019). Howover, a gene exchange/flow can provide beneficial genetic variation and may facilitate adaptation to environmental change (Li et al., 2021b; Wood et al., 2021). This is because that gene flow is expected to increase heterozygosity and reintroduce variation that can allow for masking or purging of fixed deleterious alleles (Ferrer et al., 2021; Muola et al., 2021). Unexpected, the OTc group (central populations) of O. taihangensis presented a slightly lower genetic diversity than that of OTd group (marginal populations) (Table 1). This might be related with the relative higher gene flow from OTd to OTc (Fig. 3).
Multilocus analysis resulted in significantly negativeF IS values for the inbreeding coefficient, indicating the presence of heterozygote excess for Opisthopappus species. The phenomenon of excess heterozygotes (e.g., Carapa procera ,Dioon edule , Prunusavium ) has widely been observed in other species. In general, the main reasons for the excess of heterozygotes are low number of individuals in the breeding population, overdominance, stepwise selection for homozygotes and negative assortative mating (Million, 2021; Stoeckel et al., 2006). The excess of heterozygotes implied that the species possessed a rich genetic diversity that favored the adaptation of the species to different environments (Tay Fernandez et al., 2021; Theodoridis et al., 2021). Because of the lack