Archived DNA Reveals Marine Heatwave-Associated Shifts in Fish Assemblages

Zachary Gold^{1 , 2} , Ryan P. Kelly³, Andrew Olaf Shelton², Andrew Thompson⁴, Kelly D. Goodwin⁵, Ramon Gallego², Kim Parsons², Luke R. Thompson^{5 , 6}, Dovi Kacev⁷, Paul H. Barber⁸

¹ Cooperative Institute for Climate, Ocean, & Ecosystem Studies, UW, Seattle, WA
² Northwest Fisheries Science Center, NMFS/NOAA, Seattle, WA
³ School of Marine and Environmental Affairs, UW, Seattle, WA
⁴ Southwest Fisheries Science Center, NMFS/NOAA, La Jolla, CA
⁵ Ocean Chemistry and Ecosystems Division, Atlantic Oceanographic and Meteorological Laboratory, Miami, FL
⁶ Northern Gulf Institute, Mississippi State University, Mississippi State, MS
⁷ Scripps Institution of Oceanography, UCSD, La Jolla
⁸ Department of Ecology and Evolutionary Biology, UCLA, Los Angeles, CA

Description

This page is dedicated to hosting data and code generated for the manuscript.

Included on this page is

Scripts used to Conduct Analyses

/analysis
1. CalCOFI_results_short_20210928.Rmd This script does the main analyses in the paper.
2. calcofi_metadata_20210907.Rmd This script organizes metadata.
3. Calcofi_satellite_data_2021090-7_.Rmd This script obtains SST data for analyses.
4. Calcofi_edna_vs_morphology_20210908.Rmd This script formats data for the STAN model.
5. Run_Mifish_Joint_Model.R This script runs the joint STAN model
6. Mifish_Joint_Model.stan This is the STAN model.

/decon

  1. *20210927_Calcofi_decontam_zjg.R* This script processes the "raw" *Anacapa Toolkit* output and runs decontamination to remove poorly sequenced samples and contaminant ASVs
  2. *20210927_Merge_data.Rmd* This script formats the output from decontamination and sums ASVs by taxonomy.

Data

/anacapa_output_20210507

1. *combo_q30* *Anacapa Toolkit* Output of MiFish 12S data using the *CRUX* Global Reference database from [Gold et al. 2021](https://onlinelibrary.wiley.com/doi/epdf/10.1111/1755-0998.13450       
2. *fishcard_q30* *Anacapa Toolkit* Output of MiFish 12S data using the *CRUX* Local Reference database from Gold et al. 2021.

/CalCOFI_Database_194903-201907_csv_30Apr2020

1. Directory includes metadata files

**Note** *Bottle data file needed for metadata is not uploaded due to size*

1. *larval_counts_20210305.csv* Microscopy derived morphological data
2. *mifish_library_prep.csv* Relevant sequence library preparation information needed for the joint model.
3. *20210622_species_mapping_file.csv* Species mapping file that links taxonomy obtained from morphological and molecular data.
4. *habitat_association_to_check_art.csv* Habitat association data for all species
5. Other files are various metadata and intermediate outputs needed to link data to each other.

Raw sequence inputs and large model outputs are available on Dryad. Reference databases from Gold et al. 2021 are available GitHub & Dryad

Dependencies

R (v4.0)

1. tidyverse
2. rstan
3. bayesplot
4. here
5. ncdf4
6. parsedate
7. plotdap
8. rerddap
9. sp
10. rerddapXtracto
11. devtools
12. gganimate
13. ggplot2
14. plotdap
15. lme4
16. rstanarm
17. modelr
18. tidybayes
19. heatmaply
20. wesanderson
21. dendextend
22. ggpmisc
23. vegan
24. ggdendro
25. rioja
26. cluster
27. ggrepel
28. grid
29. RColorBrewer
30. phyloseq
31. metagMisc
32. proxy
33. reshape2
34. microDecon
35. stringr
36. knitr
37. cowplot
38. reshape2
39. microDecon
40. stringr
41. ranacapa
42. plotly
43. optparse
44. fitdistrplus
45. broom
46. analyze.stuff
47. ggallin
48. patchwork

See individual packages for installation

1: Organize and Format Metadata

a. Run Calcofi_satellite_data_2021090-7_.Rmd

This script generates SST data needed for downstream analyses.

b. Run calcofi_metadata_20210907.Rmd

This script pulls relevant metadata for each CalCOFI jar analyzed in this study.

2: Decontamination

a. Run 20210927_Calcofi_decontam_zjg.R

This script imports "raw" Anacapa Toolkit data and runs a 4 step decontamination process following the methods of Kelly et al. 2018

Cleaning Process 0: Remove all Forward, Reverse, and Unmerged reads & Remove Singletons

Cleaning Process 1: Estimation of Tag-jumping or sample Cross-talk or Index hopping

Recent evidence has found that there is the potential for indexes to hop from one DNA molecule to another, leading to the incorrect sample assignment during demultiplexing Costello et al., 2018. To estimate the frequency of index hopping we included two positive controls of a non-native terrestrial taxa. To estimate the frequency of index hopping we model the composition of environmental sequences observed on the positive controls and subtract these sequences from the environmental samples run.

Cleaning Process 2: Discard PCR replicates with low number of reads

The minimum read cutoff was 30,000 reads.

Cleaning Process 3: Discard PCR (technical) replicates with unusually high dissimilarity

We removed all PCR replicates that were > 95% probability of belonging to a fit beta distribution of pairwise Bray-Curtis dissimilarities among all technical replicates.

Cleaning Process 4: Remove known lab contaminants

We removed human and pig sequences as these are known contaminants in our lab and reagents.

b. Run 20210927_Merge_data.Rmd

We sum ASV read counts by taxonomy [e.g. the reads from 4 ASVs all assigned to Northern Anchovy (Engraulis mordax) are summed together]

CalCOFI/CalCOFI_eDNA