Comparing genetic Ne reconstructions over time with long-time wolf monitoring data in two populations
This project its intended to perform some analysis to test the functionality of GONE software in especific situations see https://github.com/esrud/GONE
This repository is intended to sum up all the scripts used on the project so it's explained step by step:
In this study we used two different datasets one regarding Minnesota wolves and the other regarding Scandinavian wolves this data need some preparation before being able to run GONE, so it's important to take the next steps if it's necessary for your data.
Data from Minnesota comes from doi_10.5061_dryad.284tf__v1 this data is not compatible with GONE as a .ped file input, so it's used the script PEDconv.py on the folder AuxiliarScripts
On the same way, to prepare the map file, as the data we got needed to be converted to a map file too, it was created a script called MAPconv.py that you can find in the folder AuxiliarScripts to convert the data from the snps of the data_123801SNP.csv file that we got from Medrano, et al. (2014) using the caninehd_b dataset, we converted the file to create a usable .map file.
On the other side the data from Scandinavia was already on .ped and .map format so it's no necessary to take this step.
To sort the chromosomes by it's number (which is necessary for the data ingest in GONE) it's used the Minnesota_1_script_sorting.sh further description of the file on it's folder: Executionscripts
After that process it's requiered to limit the number of snps to 50k as GONE use that number on it's execution, in order to do that the script run it's Minnesota_2_filtering_gone_50kSNP.sh that you can find in Executionscripts also more information is provided on that folder readme.
On the final step that include the execution of GONE on 40 repetitions it's the script_subsets.sh, this script uses plink to generate subsets of less individuals in order to create the permutations necessary to run GONE 40 times with different combination of individuals, to choose the number if individuals it's used the binomial coefficient to calculate the number of possible combinations to reach at least 40 possible combinations. You can find this script on Executionscripts also more information is provided on that folder readme. This script it's also necessary to be modified to be compatible with different data sources.
On Minnesota there is an analysis made on Jupyter notebook, that it's commented and explained on every step made. It's file is called Minnesota_analysis.ipynb and it's located in the folder AnalysisScripts.
Also in Scandinavia there is an analysis made on a Jupyter notebook, in this analysis are used two datasets one with immigrants and the other one without them, this is ploted on a same graph so its convenient to process both of them in the same script. The file is located in the folder AnalysisScripts as Scandinavia_analysis.ipynb inside the file you can find further explanation.
You can find the ped and map used on gone on the folder PED_and_MAP_FILES these are the ones that are alread filtered with the plink scripts on the execution scripts. Also, in the same folder there are the ped and map files generated for Minnesota with the Auxiliar scripts.
This project is intended to be used with a linux machine, all the previous and post GONE execution could be executed on a normal desktop/laptop with enough computation power. In order to run GONE it's usual to not be able to run in a normal laptop/desktop it's irrelevant it's power, it will just not Run. We recommend to use a a cluster available on your research group, or use a cloud computation service, with docker available to run the program.
To run the auxiliar scripts it's requiered Python>3.9 and the library pandas.
To run the pre-execution scripts it's requiered to have plink and plink2 in your computer
To run the analysis scripts it's requiered to use Jupyter notebook and Python>3.9, also it's requiered to use the following libraries: matplotlib, seaborn, pandas, scipy. You can install using the following line
Anaconda
conda install pandas scipy seaborn matplotlibpip
pip install pandas scipy seaborn matplotlib