Whole-genome sequencing reveals adaptations of hairy-footed jerboas (Dipus, Dipodidae) to diverse desert environments

Abstract

Background: Environmental conditions vary among deserts across the world, spanning from hyper-arid to high-elevation deserts. However, prior genomic studies on desert adaptation have focused on desert and non-desert comparisons overlooking the complexity of conditions within deserts. Focusing on the adaptation mechanisms to diverse desert environments will advance our understanding of how species adapt to extreme desert environments. The hairy-footed jerboas are well adapted to diverse desert environments, inhabiting high-altitude arid regions, hyper-arid deserts, and semi-deserts, but the genetic basis of their adaptation to different deserts remains unknown.

Results: Here, we sequenced the whole genome of 83 hairy-footed jerboas from distinct desert zones in China to assess how they responded under contrasting conditions. Population genomics analyses reveal the existence of three species in hairy-footed jerboas distributed in China: Dipus deasyi, Dipus sagitta, and Dipus sowerbyi. Analyses of selection between high-altitude desert (elevation ≥ 3000m) and low-altitude desert (< 500m) populations identified two strongly selected genes, ATR and HIF1AN, associated with intense UV radiation and hypoxia in high-altitude environments. A number of candidate genes involved in energy and water homeostasis were detected in the comparative genomic analyses of hyper-arid desert (average annual precipitation < 70mm) and arid desert (< 200mm) populations versus semi-desert (> 360mm) populations. Hyper-arid desert animals also exhibited stronger adaptive selection in energy homeostasis, suggesting water and resource scarcity may be the main drivers of desert adaptation in hairy-footed jerboas.

Conclusions: Our study challenges the view of deserts as homogeneous environments and shows that distinct genomic adaptations can be found among desert animals depending on their habitats.

Keywords: Dipus; Energy homeostasis; High-altitude adaptation; Hyper-arid adaptation; Population genomics; Water balance.

Fig. 1

Geographic distribution and population genetic analyses of Dipus jerboas in deserts of China. A Sampling sites of 83 Dipus jerboas in China desert region covering 8 geographical populations. The elevation (m) of the study area is also visualized. B Geographic variation of the annual mean precipitation (mm). Precipitation data from 1961 to 2019 were downloaded from the WorldClim database (https://www.worldclim.org/, last accessed March 18, 2022). The 200 mm average annual precipitation line is also visualized. C Neighbor-joining (NJ) tree based on whole-genome SNPs using p-distances between individuals. The long-eared jerboa, Euchoreutes naso, is used as an outgroup. The number beside the tree nodes indicates the bootstrap value. D NJ tree based on the mitochondrial genome of all 83 Dipus individuals. E Principal components analysis of 83 Dipus individuals based on whole-genome SNPs. F Population genetic structure of the 83 Dipus jerboas inferred from the program sNMF v.1.2. The length of each color segment represents the proportion of the individual genome inferred from ancestral populations (K= 5-6, which had a low cross-validation error). See supplementary Table s1, Supplementary Material online for the abbreviations of the geographical populations and individuals

Fig. 2

Demographic history of three Dipus species of China. PSMC results for the representative individuals with high read coverage show different demographic histories of the three Dipus species with a generation time (g) of 1.5 years and a mutation rate (μ) of 6.319× 10⁻⁹ per site per generation

Fig. 3

Genomic regions with strong selective signals in the genome of Dipus sowerbyi from the high-altitude desert environment. Distribution of the A fixation indices (F_ST) and B π-ratio values calculated in 100-kb sliding windows with 50-kb steps. The horizontal blue lines above the figure represent the top 5% threshold with A F_ST > 0.51 and B π-ratio > 3.51. The candidate genes shown in A and B are closely related to high-altitude desert adaptation, with functions related to radiation response, hypoxia, DNA repair, and lipid homeostasis. C Distribution of log₂(π-ratios) and F_ST values calculated in 100-kb sliding windows with 50-kb increments between high-altitude desert group and the control low-altitude desert group. Data points in blue (corresponding to the top 5% of the empirical log₂(π-ratios) values and the top 5% of the empirical F_ST values) are genomic regions under selection in the high-altitude desert group. D Log₂(π-ratios) and F_ST values were calculated using VCFtools for each 50-kb sliding window with 25-kb increment around the gene ATR. E Allele frequencies of seven missense mutations within the ATR gene between the high-altitude desert group and the low-altitude desert group

Fig. 4

Genomic regions with strong selective signals in the genome of Dipus sowerbyi from the hyper-arid desert environment. Distribution of the A fixation indices (F_ST) and B π-ratio values calculated in 100-kb sliding windows with 50-kb steps. The horizontal blue lines above the figure represent the top 5% threshold with A F_ST > 0.18 and B π-ratio > 0.98. The candidate genes shown in A and B are closely related to hyper-arid desert adaptation, with functions related to energy metabolism and cellular homeostasis. C Distribution of log₂(π-ratios) and F_ST values calculated in 100-kb sliding windows with 50-kb increments between hyper-arid desert group and the control semi-desert group. Data points in blue (corresponding to the top 5% of the empirical log₂(π-ratios) values and the top 5% of the empirical F_ST values) are genomic regions under selection in the hyper-arid desert group. D Log₂(π-ratios) and F_ST values were calculated using VCFtools for each 50-kb sliding window with 25-kb increment around the gene DMXL2. E Allele frequencies of six missense mutations within the DMXL2 gene between hyper-arid desert group and semi-desert group