Fig. 3: GradientForest modelling genotype-climate associations and genomic offsets to future climate conditions of the two mountainous species. | Nature Communications

Fig. 3: GradientForest modelling genotype-climate associations and genomic offsets to future climate conditions of the two mountainous species.

From: The combination of genomic offset and niche modelling provides insights into climate change-driven vulnerability

Fig. 3

a, b Left, principal component analyses of gradientForest transformed climatic variables for T. elliotii (a) and P. monticolus (b). Arrows show the loadings of the top-ranked uncorrelated climatic variables. The labelled vectors of the first two principal components indicate the direction and magnitude of correlation. Right, gradientForest transformed genotype-climate association across the distribution ranges of T. elliotii (a) and P. monticolus (b). The difference in genetic composition is mapped by assigning the three principal components to the RGB colour palette according to the gradientForest manual, with the resulting colour corresponding to the expected patterns of genetic composition. c, d Left, the gradientForest predicted genomic offsets under RCP8.5 2050 in T. elliotii (c) and P. monticolus (d). Right, populations in the western parts of the distributions, i.e., the southern Tibetan zone (STZ, T. elliotii, n = 12,912 grids) and eastern Himalayan zone (EHZ, P. monticolus, n = 9224 grids), show greater genomic offsets than the populations inhabiting the eastern (LPZ&EMZ, Loess Plateau zone and east meadow zone, T. elliotii, n = 8694 grids, P. monticolus, n = 10,551 grids; WMPZ, western mountainous plateau zone, T. elliotii, n = 6902 grids, P. monticolus, n = 26,979 grids) and southern parts of the distribution ranges (SMZ, southwest mountainous zone, T. elliotii, n = 14,981 grids, P. monticolus, n = 20,275 grids). We tested genomic offsets among these groups using the two-tailed Wilcoxon rank-sum test and FDR correction for multiple comparisons. The box plots show the median (centre line) and 25th−75th percentiles (box limits). The whiskers extend to the top/bottom to the maxima and minima. Data beyond the end of the whiskers are considered outliers. White broken lines demonstrate ecological zones. See Supplementary Fig. 1 for the predicted genomic offsets under different emission scenarios and decades (RCP4.5 2050, RCP8.5 2050, RCP4.5 2070 and RCP8.5 2070).

Back to article page