Genomic divergence, local adaptation, and complex demographic history may inform management of a popular sportfish species complex

Abstract

The Neosho Bass (Micropterus velox), a former subspecies of the keystone top-predator and globally popular Smallmouth Bass (M. dolomieu), is endemic and narrowly restricted to small, clear streams of the Arkansas River Basin in the Central Interior Highlands (CIH) ecoregion, USA. Previous studies have detected some morphological, genetic, and genomic differentiation between the Neosho and Smallmouth Basses; however, the extent of neutral and adaptive divergence and patterns of intraspecific diversity are poorly understood. Furthermore, lineage diversification has likely been impacted by gene flow in some Neosho populations, which may be due to a combination of natural biogeographic processes and anthropogenic introductions. We assessed: (1) lineage divergence, (2) local directional selection (adaptive divergence), and (3) demographic history among Smallmouth Bass populations in the CIH using population genomic analyses of 50,828 single-nucleotide polymorphisms (SNPs) obtained through ddRAD-seq. Neosho and Smallmouth Bass formed monophyletic clades with 100% bootstrap support. We identified two major lineages within each species. We discovered six Neosho Bass populations (two nonadmixed and four admixed) and three nonadmixed Smallmouth Bass populations. We detected 29 SNPs putatively under directional selection in the Neosho range, suggesting populations may be locally adapted. Two populations were admixed via recent asymmetric secondary contact, perhaps after anthropogenic introduction. Two other populations were likely admixed via combinations of ancient and recent processes. These species comprise independently evolving lineages, some having experienced historical and natural admixture. These results may be critical for management of Neosho Bass as a distinct species and may aid in the conservation of other species with complex biogeographic histories.

Keywords: conservation; demography; divergence; diversity; gene flow; local adaptation.

FIGURE 1

Species geographic ranges, sampling sites, and distinct evolutionary lineages. (a) Native ranges of the Smallmouth Bass (Micropterus dolomieu; light gray) and the Neosho Bass (M. velox; dark gray), with representative illustrations. (b) Maximum‐likelihood phylogeny for putatively pure (p‐Pure) Spotted Bass, Smallmouth Bass, and Neosho Bass, with black and gray boxes at nodes indicating 100% and > 90% bootstrap support, respectively, and population structure results for K = 3, K = 4, and K = 5, with major lineages and sample sites labeled corresponding to individual samples. (c) 10‐fold cross‐validation error results for admixture analysis. (d) Sampling sites (numbered as in Table 1) within the Central Interior Highlands (CIH) for Smallmouth and Neosho Bass colored by distinct evolutionary lineages. Sites of putative admixed origin based on preliminary admixture analysis (p‐Admixed) are indicated as white circles; empty white circles indicate sites where all individuals were of putatively admixed origin, and stars indicate sites where nearly all individuals were of putatively admixed origin.

FIGURE 2

Co‐ancestry and phylogenomic relationships between Smallmouth Bass (Micropterus dolomieu) and Neosho Bass (M. velox) inferred in fineradstructure for (a) p‐Pure samples, and (b) p‐Admixed samples. We used the full non‐thinned dataset (98,659 SNPs) to generate SNP haplotypes for co‐ancestry and phylogenomic assessment. Colors within the co‐ancestry matrices reflect the extent of co‐ancestry between adjacent samples, with white and blue representing low co‐ancestry, and pink and black representing high co‐ancestry. Inferred populations are colored on the vertical at left and along the top horizontal for reference, including the hatchery‐stocked LAKE site (SKIA, navy blue), sites within the MRT (MISS, orange), sites within the WRT (WHITE, deep pink), Lee Creek and the Mulberry River (LMULB, dark green), tributaries of the middle ARB (MIDARK, sky blue), the Illinois River system (ILLI, light purple), Illinois Bayou River and Big Piney Creek, AR (BAYOU, light green), the Elk River (ELK, light brown), and tributaries of the upper ARB (UPPARK, light pink). Inset maps show geographic locations of each population; pie charts at Sites 13, 14, and 15 indicate the proportion of individuals assigned as p‐Pure (sky blue, a and b) and p‐Admixed (white, a; light purple, b). All branches on the phylogenies at left are supported with 100% of bootstrap replicates.

FIGURE 3

Discriminant Analysis of Principal Components (DAPC) results for (a) outlier (156) and (b) neutral (41,324) SNPs shared by bayescan and pcadapt for all Smallmouth Bass (Micropterus dolomieu) and Neosho Bass (M. velox) samples; (c) outlier (29) and (d) neutral (35,038) SNPs shared by bayescan and pcadapt for Neosho Bass only. (e) Geographic location of populations.

FIGURE 4

Admixture drift tree for Smallmouth Bass (Micropterus dolomieu) and Neosho Bass (M. velox) populations inferred in mixmapper. Black labels plotted on the tips of the scaffold tree with corresponding‐colored squares represent pure, unadmixed populations (as determined by a two‐way f ₃ test). Colored and shadowed branches and labels mapped onto the scaffold tree represent significantly admixed populations originating from their respective parents. The scale for branch lengths is in drift units (D) in which D is roughly equal to 2F _ST.

FIGURE 5

Best‐fitting two‐population demographic models for Smallmouth Bass (Micropterus dolomieu) and Neosho Bass (M. velox), specifically for (a) ELK and WHITE populations, (b) ILLI and SKIA populations, (c) UPPARK and WHITE populations, and (d) BAYOU and WHITE populations, generated in 𝛿a 𝛿 i. Model schematics of divergence and migration are shown in the top left‐hand inset; lengths of arrows correspond roughly to the length of time (for T parameters) or the rate of migration (for m parameters) inferred in 𝛿a 𝛿 i. The simulated 2D joint site frequency spectrum for the best‐fitting demographic model, and the 2D joint site frequency spectrum for the empirical data, are shown in the top and bottom panels of the lower right‐hand inset, respectively. Residuals representing the closeness of fit of empirical data to each model are given in Figure S14.