Ecological genetics of isolated loach populations indicate compromised adaptive potential

Abstract

Many endangered species live in fragmented and isolated populations with low genetic variability, signs of inbreeding, and small effective population sizes - all features elevating their extinction risk. The flat-headed loach (Oreonectes platycephalus), a small noemacheilid fish, is widely across southern China, but only in the headwaters of hillstreams; as a result, they are spatially isolated from conspecific populations. We surveyed single nucleotide polymorphisms in 16 Hong Kong populations of O. platycephalus to determine whether loach populations from different streams were genetically isolated from each other, showed low levels of genetic diversity, signs of inbreeding, and had small contemporary effective population sizes. Estimates of average observed heterozygosity (H_O = 0.0473), average weighted nucleotide diversity (π_w = 0.0546) and contemporary effective population sizes (N_e = 10.2 ~ 129.8) were very low, and several populations showed clear signs of inbreeding as judged from relatedness estimates. The degree of genetic differentiation among populations was very high (average F_ST = 0.668), even over short geographic distances (<1.5 km), with clear patterns of isolation by distance. These results suggest that Hong Kong populations of O. platycephalus have experienced strong genetic drift and loss of genetic variability because sea-level rise after the last glaciation reduced connectedness among paleodrainages, isolating populations in headwaters. All this, together with the fact that the levels of genetic diversity and contemporary effective population sizes within O. platycephalus populations are lower than most other freshwater fishes, suggests that they face high local extinction risk and have limited capacity for future adaptation.

Fig. 1. Study populations of flat-headed loaches.

Map of the study area showing the sampled hillstreams in the territory of Hong Kong with different palaeo-drainages indicated in color (green: Tai Mo Shan drainage - TMS; pink: West New Territories drainage - NTW; blue: South Lantau drainage - LTS; purple: Hong Kong Island drainage - HKI; orange: East New Territories drainage - NTE; gray: Other paleodrainages not included in this study) redrawn from Fyfe et al. (2000). Light blue depicts the current marine area. Fish insert at the top (left) depicts the study species, the flat-headed loach. For locality abbreviations, see Supplementary Table S1 (Photo courtesy: Chi Kit Yeung).

Fig. 2. Relatedness between individuals within each of the 16 flat-headed loach populations.

The different color boxes represent populations in different paleodrainage systems (see Fig. 1). Each row and column corresponds to a particular individual in a given population.

Fig. 3. Estimates of contemporary effective population size (N_e) in 16 loach populations as obtained with the program NeEstimator (Do et al. 2014).

Vertical lines depict 95% Parametric CIs. CIs and N_e for SHW and YSO, respectively, were not estimable. The different colors represent populations in different paleodrainage systems (see Fig. 1).

Fig. 4. The demographic history of 16 flat-headed loach populations inferred by a folded SFS model projected to NJ tree.

Different colors represent different paleodrainage systems (see Fig. 1). Red lines: the median estimated historical effective population sizes. Gray lines: 95% confidence interval of the inference. The periods of glaciation and interglaciation are highlighted with blue and orange vertical bars, respectively. The gray vertical area depicts a gap in the sedimentary record of Hong Kong.

Fig. 5. Population structure among sixteen flat-headed loach populations.

The different font colors represent populations in different paleodrainage systems (see Fig. 1). A Pairwise F_ST estimates among populations. B Population structure inferred from admixture analysis assigning individuals to the optimal K = 16 (See Supplementary Fig. S3). Each colored bar represents one individual, and colored segments correspond to different ancestral components.

Fig. 6. Isolation by distance across flat-headed loach populations.

Isolation by distance with A geographic distance (km) calculated from site coordinates (r = 0.367, p = 0.006); and B from the shortest possible waterway based on the paleodrainage system map (r = 0.207, p = 0.031). Pairwise F_ST in both plots obtained with Rousset’s (1997) correction.

Fig. 7. Results of principal component analysis (PCA) of allele frequencies projected onto the map of study area.

Note that PCA Axis 2 is depicted on the x-axis.

Fig. 8. Bubble plot illustrating the relationship between F_ST and extent of geographic sampling area for studies of freshwater fishes published between 2013 and 2023 using single nucleotide polymorphism loci (SNPs).

Bubble size indicates the number of populations and color indicates the number of SNP loci after log₁₀ transformation. The F_ST estimate for loaches from the current study is highlighted with an orange circle. For details of the underlying data, see Supplementary Table S6 and Supplementary Material. For linear models fitted into this data, see Supplementary Table S7.