ReCER_base_analysis_pipeline

These are the scripts for the ReCER base analysis which is required to be done on N drive.

This will make your metadata file for you and run an analysis.

Summary

1. set up directories, download metadata from RNR, and download dart data
2. clone this git to get scripts
3. customise 0_setup_variables.xlsx (Variables maindir, species, dataset, and raw_meta_path must be changed, others can be left default)
4. run 1_make_meta.R which makes metadata file, including automatically creating sites based on distance
5. run 2_run_analyses.R which produces summary pdf output as well as specific plots and tables

1. Set up directories and data

Make a directory that will contain all of your data and analyses. This can be a species name (e.g. acacia_suaveolens) or another relevant descriptor (e.g. red_box_gums). This is your main directory.

Within the main directory, create a subdirectory. Usually this would be a combination of the genus and species names (e.g. Acacia suaveolens -> AcacSuav). This is the working directory, which will contain all of your data and outputs.

Within the working directory, make dart_raw and meta subdirectories. Save your DArT SNP mapping csv to dart_raw.

Download the relevant RnR export into meta. This export should contain all of the samples you wish to analyse. It's okay if it has irrelevant samples, they will be removed.

Now your directories should look like this:

--acacia_suaveolens
  `--AcacSuav
    |--dart_raw
    |   `-Report_DAca20-5040_SNP_mapping_2.csv
    `--meta
        `-Tissue-2023-11-02_135351.csv

2. Clone the files in this git

You need to get a copy of the files in this github into your main directory.

To do this you can either:

1. git clone if you have a mac (git clone git@github.com:eilishmcmaster/ReCER_base_analysis_pipeline.git)
2. Manually download zip (go to Code >Download zip in top right of https://github.com/eilishmcmaster/ReCER_base_analysis_pipeline)

When you're done it your directory should look like this:

--acacia_suaveolens
  |-- 0_setup_variables.xlsx
  |-- 1_make_meta.R
  |-- 2_run_analyses.R
  `--AcacSuav
    |--dart_raw
    |   `-Report_DAca20-5040_SNP_mapping_2.csv
    `--meta
        `-Tissue-2023-11-02_135351.csv

3. Customise analysis parameters

To set up your analysis, open 0_setup_variables.xlsx in excel. This file contains input variables that the R scripts use to do your analyses.

Here is an example of what 0_setup_variables.xlsx could look like:

Make sure that you have entered the correct values for maindir, species, dataset, and raw_meta_path -- these will be specific to your project. The rest of the variables are generic, so you can change them or not.

! IMPORTANT !

The variables species_col_name and site_col_name are very important for your analysis. You cannot have multiple species_col_name conditions with the same site_col_name conditions i.e. sites cannot contain multiple species.

In your analyses, species_col_name specifies the variable containing species/genetic groups. This is used for visualisation (e.g. map of different species_col_name groups) and diversity analyses (i.e. in diversity analysis, your loci are filtered to be relevant by species_col_name groups). The default workflow will use the species names in RNR (column sp in meta file produced by 1_make_meta.R).

• If you have multiple species in your dataset and you're confident about their IDs, use the default species_col_name = sp
• If you are not confident about species in your dataset, use species_col_name = none
• If you have custom genetic groups or species, use your custom column name in you metadata e.g. species_col_name = genetic_groups_custom (you will have to add your custom columns to the metadata AFTER running 1_make_meta.R

In your analyses, site_col_name specifies the variable used for sites or populations. By default this is set to the site variable, which is created by 1_make_meta.R by grouping geographically proximate samples (i.e. samples<1km apart are in the same site). If you would prefer to use a different metadata variable as site_col_name, specify this in 0_setup_variables.xlsx.

4. Generate metadata

Once your 0_setup_variables.xlsx is complete, open and run 1_make_meta.R. This script turns your RNR data into RRtools ready metadata. You shouldn't have to modify anything within the script, as variables are supplied through 0_setup_variables.xlsx.

1_make_meta.R main functions include:

• creates site column by determining which samples are <1km apart and grouping them
• creates sp column from RNR accepted species
• If your RNR data contains parent and seedling information, it processes that into tissue, mother, and family columns

It also creates additional subdirectories that your further analysis needs.

! IMPORTANT !

If you already have a metadata file that contains sample, site, lat, long etc, you do not have to use the metadata file produced here. You should still run this to make the required directories, but you can either delete the output meta or comment out line 154. Also make sure you specify your variables in 0_setup_variables.xlsx. If you want to exclude samples, simply leave the site_col_name column blank for that sample.

5. Run analysis

Once your metadata has been made you can run 2_run_analyses.R. You shouldn't have to modify anything within the script, as variables are supplied through 0_setup_variables.xlsx.

This will produce a directory within your working directory that contains figures, tables, and other outputs. The final pdf including most analyses (except splitstree and kinship) will be called something like AcacSuav_combined_outputs.pdf in your outputs folder.

Kinship heatmaps are also available in outputs/plots/ directory, but are usually too large to sensibly include in the combined outputs pdf. The splitstree file is made by this script but it has to be opened and saved using the Splitstree software before it can be re-imported and visualised in R. The code for this is commented out in 2_run_analyses.R.