Fine-scale population structure and evidence for local adaptation in Australian giant black tiger shrimp (Penaeus monodon) using SNP analysis

Abstract

Background: Restrictions to gene flow, genetic drift, and divergent selection associated with different environments are significant drivers of genetic differentiation. The black tiger shrimp (Penaeus monodon), is widely distributed throughout the Indian and Pacific Oceans including along the western, northern and eastern coastline of Australia, where it is an important aquaculture and fishery species. Understanding the genetic structure and the influence of environmental factors leading to adaptive differences among populations of this species is important for farm genetic improvement programs and sustainable fisheries management.

Results: Based on 278 individuals obtained from seven geographically disparate Australian locations, 10,624 high-quality SNP loci were used to characterize genetic diversity, population structure, genetic connectivity, and adaptive divergence. Significant population structure and differentiation were revealed among wild populations (average F_ST = 0.001-0.107; p < 0.05). Eighty-nine putatively outlier SNPs were identified to be potentially associated with environmental variables by using both population differentiation (BayeScan and PCAdapt) and environmental association (redundancy analysis and latent factor mixed model) analysis methods. Clear population structure with similar spatial patterns were observed in both neutral and outlier markers with three genetically distinct groups identified (north Queensland, Northern Territory, and Western Australia). Redundancy, partial redundancy, and multiple regression on distance matrices analyses revealed that both geographical distance and environmental factors interact to generate the structure observed across Australian P. monodon populations.

Conclusion: This study provides new insights on genetic population structure of Australian P. monodon in the face of environmental changes, which can be used to advance sustainable fisheries management and aquaculture breeding programs.

Keywords: Aquaculture; Genotype–environment interaction; Population genetics; Prawn; Functional annotation.

Fig. 1

Population structure of 278 individuals of Penaeus monodon samples using 10,624 SNPs (BB: Bramston Beach, EB: Etty Bay, Townsville: TSV, Gulf of Carpentaria: GC, Joseph Bonaparte Gulf: JBG, Tiwi Island: TIW, and Nickol Bay: NKB). a Discriminant Analysis of Principal Components (DAPC) scatterplot and b an individual density plot on the first discriminant function created through the R package ‘adegenet ‘; and c and d Population networks constructed using the Netview P v.1.0 pipeline at kNN = 30 and 60. Dots represent individuals, whereas colored ellipses correspond to sampling origin

Fig. 2

Venn diagram showing the total number of putative SNP significantly associated with at least one environmental variable. The total number of SNP is reported in each panel. The 89 outlier SNPs on the intersections (shaded area) were retained for further analysis

Fig. 3

Principal components analysis based allele frequencies for (a) 10,535 neutral SNPs loci and (b) 89 outlier SNPs loci (BB: Bramston Beach, EB: Etty Bay, Townsville: TSV, Gulf of Carpentaria: GC, Joseph Bonaparte Gulf: JBG, Tiwi Island: TIW, and Nickol Bay: NKB)

Fig. 4

Redundancy analysis on (a) 10,535 neutral loci and (b) 89 outlier SNPs allele frequencies in seven populations of Penaeus monodon in Australia. Explanatory variables (arrows) were RDA axes retained as important variable selection accounting for genetic variation (Env_2: surface temperature maximum, Env_3: surface temperature minimum, Env_5: surface phytoplankton mean, Env_7: benthic current velocity mean, and Env_1: surface salinity mean)

Fig. 5

Map showing the seven localities where 283 wild Penaeus monodon samples were collected (BB: Bramston Beach, EB: Etty Bay, Townsville: TSV, Gulf of Carpentaria: GC, Joseph Bonaparte Gulf: JBG, Tiwi Island: TIW, and Nickol Bay: NKB). The map was created using R package ‘ggplot2’ version 3.2.1 (https://github.com/tidyverse/ggplot2) [85] and R package ‘ozmaps’ version 0.3.6 (https://github.com/mdsumner/ozmaps) [86]