Research ArticleMOLECULAR BIOLOGY

Species-specific molecular responses of wild coral reef fishes during a marine heatwave

See allHide authors and affiliations

Science Advances  18 Mar 2020:
Vol. 6, no. 12, eaay3423
DOI: 10.1126/sciadv.aay3423
  • Fig. 1 Summary of the sampling design.

    (A) Five fish species analyzed in this study. (B) Temperature records for the sampling location from 2010 to 2016. Dots are daily averages in 2015–2016. The trendlines representing the 5 years before the heatwave, as well as 2015–2016, are based on the t-based approximation of the daily temperature averages. The circles at the bottom represent the species collected in that month (i.e., only A. polyacanthus and P. moluccensis were collected in February). Average temperatures for each month were as follows: December, 27.81°C (SD ±0.45°C); February, 29.55°C (SD ±0.69°C); March, 29.70°C (SD ±0.61°C); July, 24.96°C (SD ±0.28°C). (C) Schematic representation of the sample processing from liver extraction to gene expression analysis. RNASeq, RNA sequencing.

  • Fig. 2 Molecular response to the 2015–2016 marine heatwave in five coral reef fishes.

    (A) Clustering analysis across all biological replicates revealing samples grouped tightly by species and collection months represented by different colors. For all analyses, comparisons with samples from February were only possible for damselfishes A. polyacanthus and P. moluccensis. (B) Number of DEGs between samples collected at different time points of the heatwave. (CQUI, C. quinquelineatus; ODOE, O. doederleini; OCYA, O. cyanosoma; APOLY, A. polyacanthus; PMOL, P. moluccensis); N/A, not available.

  • Fig. 3 Comparison of species functional responses across the heatwave.

    (A) Principal Component Analysis of relative expression of KEGG functional pathways underlying the weighted gene correlation network analysis. (B) Heatmap representing the number of DEGs associated with specific KEGG pathways on 10 categories associated with heat stress, across different collection times for each species. Red colors represent higher numbers of DEGs in a KEGG pathway, while blue colors represent few DEGs. For (A) and (B), “Initial exposure” represents the comparison between fish collected in December and February; “Dec vs. Mar” is the comparison of samples from December and March; “Prolonged exposure” represents the comparison between February and March; and “Warmest to coldest” is the comparison between March and July.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/12/eaay3423/DC1

    Fig. S1. Number of transcripts obtained after de novo transcriptome assembly with Trinity (gray) and retained transcripts after predicting open reading frames with TransDecoder (black).

    Fig. S2. Normalized read counts for six of the genes that showed divergent expression (i.e., larger differences between species) in the analysis of EVE.

    Fig. S3. Venn diagram showing the overlap of differential expressed genes in the pairwise comparisons between December and March and December versus July for the five species of interest.

    Fig. S4. Module-trait correlation matrix for all species showing correlation value with P value in brackets for each module and trait.

    Fig. S5. Module-trait correlation matrix for A. polyacanthus and P. moluccensis (damselfish species).

    Fig. S6. Module-trait correlation matrix for specimen of all three cardinalfish species (C. quinquelineatus, O. doederleini, and O. cyanosoma).

    Fig. S7. Principal Coordinate Analysis for damselfishes, based on liver gene expression.

    Fig. S8. Eigenvalue of expression data for each biological replicate in the red module for cardinalfishes.

    Fig. S9. Heatmaps for the expression of heat-shock proteins in damselfishes.

    Fig. S10. Eigenvalue of expression data for each biological replicate in the green module for damselfishes.

    Fig. S11. Eigenvalue of expression data for each biological replicate in the red module for damselfishes.

    Fig. S12. Transcriptome completeness estimates for all five species with BUSCO.

    Fig. S13. ML tree showing the relationship between the five species analyzed in this study, generated with RAXML.

    Data file S1. Final list of orthologous genes shared across the five sampled species and their corresponding annotation.

    Data file S2. Significantly diverged genes (i.e. higher difference between species) in the Expression Variance and Evolution (EVE) analysis for all samples collected in our study, and their corresponding annotation.

    Data file S3. Enriched Gene Ontology Categories for the significantly diverged genes (i.e. higher difference between species) in the Expression Variance and Evolution (EVE) analysis.

    Data file S4. Genes with significant plasticity (i.e. higher difference within individuals of a species) in the Expression Variance and Evolution (EVE) analysis, for all samples collected in our study.

    Data file S5. Differentially expressed genes (DEGs) between fish collected at different months for the five focal species of the study. The dashes represent comparisons that were not possible, as samples from February for three of the species (C. quinquelineatus, O. cyanosoma, O. doederleini) were not available.

    Data file S6. Functional enrichment for Acanthochromis polyacanthus, based on the DEGs between samples from December and February.

    Data file S7. Functional enrichment of GO categories for Pomacentrus moluccensis, based on the DEGs between samples from December and February.

    Data file S8. Overlap in the enriched Gene Onthology categories based on the differentially expressed genes of A. polycanthus and P. moluccensis between samples from December and March.

    Data file S9. Functional enrichment of GO terms for the two species of damselfishes, based on genes belonging to the correlated network “BLUE”.

    Data file S10. Functional enrichment of GO terms for the two species of damselfishes, based on genes belonging to the correlated network “YELLOW”.

    Data file S11. Significantly enriched Gene Ontology terms between December and March for O. cyanosoma, O. doederleini and C. quinquelineatus, based on the analysis of differential expression.

    Data file S12. Overlapping differentially expressed genes between the three species of cardinalfishes when comparing samples collected in December and March.

    Data file S13. Significantly enriched GO terms for the “CYAN” gene module, for March samples of three species of cardinalfishes.

    Data file S14. Significantly enriched Gene Ontology processes for the “RED” gene module for March samples of three species of cardinalfishes.

    Data file S15. Average Resting Oxygen consumption (MO2Rest), Standard Deviation (SD in parenthesis), number of samples and temperature quotients (Q10), from studies by Nilsson et al. (2009) and Rummer et al. (2014; for C. quinquelineatus). Estimates for four of the species correspond to Lizard Island populations (i.e. same location as genetic analyses in the main text), while the samples of C. quinquelineatus correspond to Nago Island, Papua New Guinea. Only measures between 29°C and 31°C are presented in this comparison, as these resemble the temperatures during the 2016 heatwave. The values of Q10 were calculated with the equation Q10 = (MO2Rest2/MO2Rest1)[10/T2-T1].

    Data file S16. Overlap in the enriched Gene Onthology categories based on the differentially expressed genes of A. polyacanthus and P. moluccensis between samples collected in February and March.

    Data file S17. Heat shock proteins that were differentially expressed in the comparison of samples from February and March for A. polyacanthus and P. moluccensis.

    Data file S18. Significantly enriched GO terms for the “GREEN” gene module, corresponding to samples of damselfishes collected in March.

    Data file S19. Overlapping differentially expressed genes that were exclusively found in the comparisons of prolonged exposure (February vs. March) in A. polyacanthus and P. moluccensis.

    Data file S20. Significantly enriched GO terms for the “RED” gene module, resulting from the contrast of damselfishes collected in March and July.

    Data file S21. Significantly enriched GO terms for the “BLACK” gene module, corresponding to samples of cardinalfishes collected in March.

    Data file S22. Differentially expressed genes between samples of March and July for O. cyanosoma (OCYA) and O. doederleini (ODOE).

    Data file S23. Differentially expressed genes between samples of March and July of C. quinquelineatus.

    Data file S24. Common differentially expressed genes across all five species, when comparing March and July samples.

    Data file S25. Table of collected individuals, dates and fish sizes.

    Data file S26. Number of reads per individual before and after trimming and decontamination.

    Data file S27. Statistics of de novo assemblies of transcriptomes determined with BUSCO.

    Data file S28. Results from the quality assessment using the program Transrate, for contigs assembled with Trinity and summarized with Transdecoder (A) and the orthologous sequences (B).

    Data file S29. Representative transcript for each species in our study, for all orthologous genes.

    Data file S30. Number of transcripts in WGCNA modules.

  • Supplementary Materials

    The PDF file includes:

    • Fig. S1. Number of transcripts obtained after de novo transcriptome assembly with Trinity (gray) and retained transcripts after predicting open reading frames with TransDecoder (black).
    • Fig. S2. Normalized read counts for six of the genes that showed divergent expression (i.e., larger differences between species) in the analysis of EVE.
    • Fig. S3. Venn diagram showing the overlap of differential expressed genes in the pairwise comparisons between December and March and December versus July for the five species of interest.
    • Fig. S4. Module-trait correlation matrix for all species showing correlation value with P value in brackets for each module and trait.
    • Fig. S5. Module-trait correlation matrix for A. polyacanthus and P. moluccensis (damselfish species).
    • Fig. S6. Module-trait correlation matrix for specimen of all three cardinalfish species (C. quinquelineatus, O. doederleini, and O. cyanosoma).
    • Fig. S7. Principal Coordinate Analysis for damselfishes, based on liver gene expression.
    • Fig. S8. Eigenvalue of expression data for each biological replicate in the red module for cardinalfishes.
    • Fig. S9. Heatmaps for the expression of heat-shock proteins in damselfishes.
    • Fig. S10. Eigenvalue of expression data for each biological replicate in the green module for damselfishes.
    • Fig. S11. Eigenvalue of expression data for each biological replicate in the red module for damselfishes.
    • Fig. S12. Transcriptome completeness estimates for all five species with BUSCO.
    • Fig. S13. ML tree showing the relationship between the five species analyzed in this study, generated with RAXML.

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • Data file S1 (Microsoft Excel format). Final list of orthologous genes shared across the five sampled species and their corresponding annotation.
    • Data file S2 (Microsoft Excel format). Significantly diverged genes (i.e. higher difference between species) in the Expression Variance and Evolution (EVE) analysis for all samples collected in our study, and their corresponding annotation.
    • Data file S3 (Microsoft Excel format). Enriched Gene Ontology Categories for the significantly diverged genes (i.e. higher difference between species) in the Expression Variance and Evolution (EVE) analysis.
    • Data file S4 (Microsoft Excel format). Genes with significant plasticity (i.e. higher difference within individuals of a species) in the Expression Variance and Evolution (EVE) analysis, for all samples collected in our study.
    • Data file S5 (Microsoft Excel format). Differentially expressed genes (DEGs) between fish collected at different months for the five focal species of the study. The dashes represent comparisons that were not possible, as samples from February for three of the species (C. quinquelineatus, O. cyanosoma, O. doederleini) were not available.
    • Data file S6 (Microsoft Excel format). Functional enrichment for Acanthochromis polyacanthus, based on the DEGs between samples from December and February.
    • Data file S7 (Microsoft Excel format). Functional enrichment of GO categories for Pomacentrus moluccensis, based on the DEGs between samples from December and February.
    • Data file S8 (Microsoft Excel format). Overlap in the enriched Gene Onthology categories based on the differentially expressed genes of A. polycanthus and P. moluccensis between samples from December and March.
    • Data file S9 (Microsoft Excel format). Functional enrichment of GO terms for the two species of damselfishes, based on genes belonging to the correlated network “BLUE”.
    • Data file S10 (Microsoft Excel format). Functional enrichment of GO terms for the two species of damselfishes, based on genes belonging to the correlated network “YELLOW”.
    • Data file S11 (Microsoft Excel format). Significantly enriched Gene Ontology terms between December and March for O. cyanosoma, O. doederleini and C. quinquelineatus, based on the analysis of differential expression.
    • Data file S12 (Microsoft Excel format). Overlapping differentially expressed genes between the three species of cardinalfishes when comparing samples collected in December and March.
    • Data file S13 (Microsoft Excel format). Significantly enriched GO terms for the “CYAN” gene module, for March samples of three species of cardinalfishes.
    • Data file S14 (Microsoft Excel format). Significantly enriched Gene Ontology processes for the “RED” gene module for March samples of three species of cardinalfishes.
    • Data file S15 (Microsoft Excel format). Average Resting Oxygen consumption (MO2Rest), Standard Deviation (SD in parenthesis), number of samples and temperature quotients (Q10), from studies by Nilsson et al. (2009) and Rummer et al. (2014; for C. quinquelineatus). Estimates for four of the species correspond to Lizard Island populations (i.e. same location as genetic analyses in the main text), while the samples of C. quinquelineatus correspond to Nago Island, Papua New Guinea. Only measures between 29°C and 31°C are presented in this comparison, as these resemble the temperatures during the 2016 heatwave. The values of Q10 were calculated with the equation Q10 = (MO2Rest2/MO2Rest1)10/T2-T1.
    • Data file S16 (Microsoft Excel format). Overlap in the enriched Gene Onthology categories based on the differentially expressed genes of A. polyacanthus and P. moluccensis between samples collected in February and March.
    • Data file S17 (Microsoft Excel format). Heat shock proteins that were differentially expressed in the comparison of samples from February and March for A. polyacanthus and P. moluccensis.
    • Data file S18 (Microsoft Excel format). Significantly enriched GO terms for the “GREEN” gene module, corresponding to samples of damselfishes collected in March.
    • Data file S19 (Microsoft Excel format). Overlapping differentially expressed genes that were exclusively found in the comparisons of prolonged exposure (February vs. March) in A. polyacanthus and P. moluccensis.
    • Data file S20 (Microsoft Excel format). Significantly enriched GO terms for the “RED” gene module, resulting from the contrast of damselfishes collected in March and July.
    • Data file S21 (Microsoft Excel format). Significantly enriched GO terms for the “BLACK” gene module, corresponding to samples of cardinalfishes collected in March.
    • Data file S22 (Microsoft Excel format). Differentially expressed genes between samples of March and July for O. cyanosoma (OCYA) and O. doederleini (ODOE).
    • Data file S23 (Microsoft Excel format). Differentially expressed genes between samples of March and July of C. quinquelineatus.
    • Data file S24 (Microsoft Excel format). Common differentially expressed genes across all five species, when comparing March and July samples.
    • Data file S25 (Microsoft Excel format). Table of collected individuals, dates and fish sizes.
    • Data file S26 (Microsoft Excel format). Number of reads per individual before and after trimming and decontamination.
    • Data file S27 (Microsoft Excel format). Statistics of de novo assemblies of transcriptomes determined with BUSCO.
    • Data file S28 (Microsoft Excel format). Results from the quality assessment using the program Transrate, for contigs assembled with Trinity and summarized with Transdecoder (A) and the orthologous sequences (B).
    • Data file S29 (Microsoft Excel format). Representative transcript for each species in our study, for all orthologous genes.
    • Data file S30 (Microsoft Excel format). Number of transcripts in WGCNA modules.

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article