Utremi

This code is for performing ancestral niche reconstruction according to the histogram method discussed in Folk et al. 2022 (reference at the bottom of this page) and reproducing analyses from the same paper.

The quoted code should be run in bash. Ensure that paths of output from each step are not being messed with. Some paths are hard-coded in scripts and should be changed as noted. PNO files should be in a directory called "pnos", and the path to mcmc output should be updated in date_histograms_from_mcmctree.r.

Setup

If you want to run this pipeline yourself, clone the repository such that your directory setup is as follows:

Directory/   
├── data # Replace with your own data
|	├── astral_mcmctree_rep3.tre (tree file - dated)
|	├── mcmc.txt    
|	└──...   
├── pnos # Replace with your own data
|	├──
|	└── ...   
├── results # These are previously calculated results
|	└── 
├── scripts # From repository
|	├──  dating_histograms.csv (from date_histograms_from_mcmctree.r)   
|	├──   annotate_maximum_density.py   
|	├──   trim_sum_and_normalize_projections.py   
|	├──   projections_binned_ancestralreconstruction.py   
|	├──   weight_projection_by_date_probability.py   
└─ 	└──   date_histograms_from_mcmctree.r   
├── comparision
|	├──  phyloclim
|	     ├──  01_phyloclim_run.R
|	     ├── data
|	         ├── PNO
|               └─── ... 
└─           └── astral_mcmctree_rep3.tre

To duplicate paper results, run as follows without changes. The only thing missing in the path is paleoclimatic data (see Dryad).

Subfolders

data

Tree and MCMCtree output.

occurrences

Occurrence files in the expected format for pno_calc.

pnos

PNO files in the expected format.

scripts

Python and R scripts.

results

ancestral_reconstruction folders contain paper results comprising trees with the ancestral reconstruction (in BEAST-style format) and plots of both extant and ancestral reconstruction histograms. There is a folder for each of the ASTRAL and concatenation topologies. Plots of histograms are blue for the ML result, and red and green representing +- standard error.

ancestral_projection folders contain results comprising an animated GIF showing trends in habitat suitability from the Pleistocene to present based on mean annual temperature. There is a folder for each of the ASTRAL and concatenation topologies.

comparison

Comparsion with previous published methods including phyloclim and ambitus.

Dependencies

The python scripts have various dependencies that can be queried by attempting to run the scripts; full installation instructions are only given for BiotaphyPy below. Most important is a full working install of GDAL (both the Python library and the executables) that is in the path. GDAL can be challenging to install on some systems; for Ubuntu systems apt is recommended and homebrew or miniconda for OS-X. The Python bindings must also be available; e.g., typing gdal_translate in the shell should execute. A further incomplete list is:

pip3 install rasterio
pip3 install dendropy

Clone the repository

git clone https://github.com/ryanafolk/utremi
cd utremi

Set up directory structure

mkdir ancestral_reconstruction_astral

mkdir ancestral_projection_astral

Ancestral niche analyses

1. Install BiotaphyPy

cd ancestral_reconstruction_astral
git clone https://github.com/biotaphy/BiotaPhyPy
cd BiotaPhyPy/
python setup.py install
cd ..

Ensure the cloned repository remains in the working directory for the next step. If you receive errors, try running the setup.py line with "python3" instead of "python."

2. Run ancestral niche reconstruction in a loop, one iteration per variable

for f in ./../pnos/*.dropped; do
    g=$( echo ${f} | sed 's/.*\///g' | sed 's/\..*//g' )
    echo ${g}
    mkdir plots_${g}
    python3 BiotaPhyPy/biotaphy/tools/ancestral_distribution.py ./../data/astral_mcmctree_rep3.tre newick ${f} csv out_${g}.tre nexus -c out_table_${g}.txt -p plots_${g}
    done

If you get module import errors at this step, there is something wrong with the path in your BiotaPhyPy install. This seems to only affect certain systems, and I am still working on it.

3. Annotate trees by the single bin value with maximum probability density

This is used for color plotting. In the case of ties one is arbitrarily taken.

for f in out_*.tre; do
    g=`echo ${f} | sed 's/\.tre//g'`
    python3 ./../scripts/annotate_maximum_density.py ${f} ${g}.maxdensity.tre
    done

Downstream analyses

1. Generate histograms representing dating uncertainty

Generates dating histograms from MCMCtree output. Note the path to the mcmc.txt file output by this package is hard-coded on line 2 and should be updated as appropriate.

cd ..
cd ancestral_projection_astral
R CMD BATCH ./../scripts/date_histograms_from_mcmctree.r 
# A file called dating_histograms.csv should appear in the working directory

2. Perform geographic range projections on paleoclimate data

python3 ./../scripts/projections_binned_ancestralreconstruction.py ./../ancestral_reconstruction_astral/out_table_BIOCLIM_1.txt ./BIOCLIM_1 -l /mnt/Botbot/heuchera_research/Heuchera_complete_project/ShellyFinal_NovemberLayers_2019/bio1_final/*.tif
mkdir projected
mv *tifout*.tif ./projected/
# There should now be lots of TIFs in a folder called projected

3. Weight projections by posterior probability of occurrence date

File paths should reflect output of #1 and #2 above and should not be messed with just yet.

python3 ./../scripts/weight_projection_by_date_probability.py dating_histograms.csv 
# There should now be yet more TIFs in a folder called projected_weighted

4. Trim the projections to the study area, normalize histogram area, and combine across species.

mkdir combined_normalized
for i in `ls projected_weighted/*.tif | sed 's/.*_//g' | sed 's/\.tif//g' | sort | uniq`; do
    python3 ./../scripts/trim_sum_and_normalize_projections.py ./combined_normalized/${i}.combined.tif -178.2 6.6 -49.0 83.3 --rasters `ls projected_weighted/*_${i}.*tif`
    done

We end up with one TIF per time period, representing joint probabilities of occurrence for all species.

All scripts contain usage examples in header comments.

Associated Publication

Folk RA*, Gaynor ML*, Engle-Wrye NJ, O’Meara BC, Soltis PS, Soltis DS, Guralnick RP, Smith SA, Grady CJ, Okuyama Y. 2023. Identifying climatic drivers of hybridization with a new ancestral niche reconstruction method. Systematic Biology, syad018. doi: 10.1093/sysbio/syad018.

Related Repositories

mgaynor1/Paleogenerate
ryanafolk/spatial_utilities