.

GCMP Global Disease

This project hold the workflow for the Global Coral Microbiome Project's global comparison of coral disease susceptibility and microbiome structure. This project combines microbiome, phylogenetic and trait data to test long-term correlations between coral disease and microbiome structure.

Project Organization:

Analysis Folders

The analysis folder will be organized into subdirectories by analytical step. Each analytical step is required to have 3 subfolders: input output and procedure

• analysis/
  • organelle_removal/
    • input/
    • output/
    • procedure/

input folder: Typically, necessary data files cannot be checked in to GitHub. Therefore, scripts at each analytical step should download the files they need as a first step (e.g. with wget or curl).

output folder: The output folder holds all results of the analysis. All files in output should be able to be regenerated (deleted and then recreated) from the data in input using the scripts in procedure. This is essential to ensure outputs stay up-to-date if upstream QC steps change (which is common).

procedure:. The procedure should be written using Jupyter Notebooks when possible, but R markdown or bash scripts may be more convenient for some steps. If there is more than one procedure file, please create an index file that explains what order to run them in and what each is doing.

Workflow

Each step is labelled with output that is used in later steps, and products that are used in Figures or Supp. Figures in the paper To recreate the analysis we will run the procedures in the following analysis folders in order:

1. coral_disease_data

   • input: raw data from the FRRP, HICORDIS and Lamb et al, disease datasets; the Huang & Roy coral phylogeny
   • output: merged_disease_table.tsv
   • products: disease prevalence graphs

2. metadata

   • input: GCMP_EMP_map_r28_no_empty_samples.txt
   • output: one_hot_encoding_metadata.tsv

3. organelle_removal

   • input: GCMP_EMP_map_r28_no_empty_samples.txt -- the sample metadata all.seqs.fa -- the sequences from QIITA all.biom -- the feature table from QIITA
   • output: silva_metaxa2_reference_taxonomy.qza -- taxonomic annotations that better detect mitochondrial sequences GCMP_seqs.qza effects_of_rarefaction/ feature_table_silva_metaxa_2_all.qza

4. phylogeny_insertion

   • input: all.seqs.fa (large reference phylogenies are downloaded automatically by the script)

   • output: insertion-tree_GCMP.qza

     insertion-tree_SILVA_GCMP.qza

5. decontamination

   • input: feature_table_silva_metaxa_2_all.qza insertion-tree_SILVA_GCMP.qza silva_metaxa2_reference_taxonomy.qza GCMP_EMP_map_r28_no_empty_samples.txt --currently the # in #SampleID is removed manually in this version
   • output: physeq.noncton-rooted-tree.qza -- the tree of microbes, excluding contaminants physeq.noncont-feature-table.qza

6. PICRUSt2

   • inputs: physeq.noncont-feature-table.qza GCMP_seqs.qza
   • output: EC_metagenome.qza KO_metagenome.qza pathway_abundance.qza

7. core_analysis

   • input: GCMP_EMP_map_r28_no_empty_samples.txt insertion-tree_silva_GCMP.qza Copied via the setup_input_data.py script: physeq.noncton-rooted-tree.qza --> physeq_rooted_tree.qza physeq.noncont-feature-table.qza --> physeq_feature_table.qza physeq.noncont_sample-metadata.txt --> physeq_metadata.txt physeq.noncont_tax.txt --> physeq_taxonomy.txt
   • output: alpha_diversity_phylogenetic/ -- faith's pd .qza and .qzv files beta_diversity_phylogenetic/ -- weighted and unweighted UniFrac bdeta diversity distance matrix .qza files adiv_trait_table_all.txt -- a tsv file with alpha diversity and most abundant taxa calculated separately for each species (by Huang Roy tree name) adiv_trait_table_mucus.txt -- same but only for mucus adiv_trait_table_tissue.txt -- same but only for tissue adiv_trait_table_skeleton.txt -- same but only for skeleton

8. coral_disease_vs_adiv - input: - output: - products: phytools contmap plots of adiv (rightwards) vs. disease (leftwards) phylomorphospace plots for adiv vs. disease PIC results for corelations of adiv vs. disease PIC plots for correlations of adiv vs. disease heatmap of adiv vs. disease

9. GDM

   • input: combined_meta.tsv --> merged by the R script from one_hot_encoding_metadata.tsv and GCMP_EMP_map_r28_no_empty_samples.txt physeq.noncton-rooted-tree.qza --> physeq_rooted_tree.qza physeq.noncont-feature-table.qza --> physeq_feature_table.qza physeq.noncont_tax.txt --> physeq_taxonomy.txt
   • output:
   • products: