In this project, the Genealogy Registry of the Genetic Biobank of the Faroe Islands is used to select individuals from the FarGen cohort (www.fargen.fo) for whole-genome sequencing. The wgs_selection.html file is a rendered R notebook that shows how the samples were selected. Briefly, we read in a genealogy of the FarGen cohort, and calculated the kinship coefficients. We then explored population structure of this cohort before selecting a number of individuals to best represent the general population.
The dialect_map.pdf file contains some information about how the the population was divided into regions in order to explore population structure.
Before the WGS selection R notebook was run, there were some steps to produce the genealogies and associated data sets. These are shown below. The raw data used in this analysis is highly confidential, so only results are shown. The analysis was done on FarGen's High-Throughput Computational Cluster.
The AEBS repository is used to read data from exported data from the Genealogy Registry, and to reconstruct the genealogy of the individuals in the Fargen cohort. The AEBS repository can be found here: https://github.com/olavurmortensen/aebs
First add the AEBS scripts to path:
cd aebs
source setup.sh
cd ..
Clean up the raw GEDCOM file:
ged_cleanup.sh data/ged/aebs_export_2020-06-18/aebs_export_2020-06-18.ged data/ged/cleaned.ged
Extract the relevant fields from the cleaned GEDCOM file and write it to a CSV:
ged2csv data/ged/cleaned.ged data/trees/aebs.csv
Before we can reconstruct the FarGen genealogy, we must obtain the IDs of the individuals. To do that, we need the P-number of the FarGen individuals, and we shall match them with the P-numbers in the GEDCOM file.
Below we extract the REFN and RIN fields from the cleaned GEDCOM file. The REFN field contains the P-number.
ged2csv.py data/ged/cleaned.ged data/ids/aebs_ids.csv
We obtain information about the FarGen individual from the Progeny database. The data/ids/fargen_ids.csv file from Progeny contains the FarGen individual IDs FN and their P-numbers. Below we match FarGen individuals with individuals in the genealogy registry, and obtain their RIN and FN IDs in a CSV, and only their RIN in a text file.
python match_ids.py data/ids/fargen_ids.csv data/ids/aebs_ids.csv fargen_rin_samplename.csv fargen_rin.txt
Now we can reconstruct the FarGen genealogy:
python lineages.py --csv data/trees/aebs.csv --ind data/ids/fargen_rin.txt --out data/trees/fargen.csv
The data/placenames/placename_regions.csv file contains the names of the 200 most common birth places in the genealogy registry. The placenames are cleaned up with a function in the data/placenames/clean_placename.py script, to reduce the number of unique placenames.
Get a list of the 200 most common birth places:
python data/placenames/birth_locations_cleanup.py data/trees/aebs.csv data/placenames/placename_regions.csv
Then I have manually edited the data/placenames/placename_regions.csv file, adding the "region" column, so that each birth place is associated with a geographical region (1-6).
Then make a list of RIN and region, so that each individual is attributed a regional birth place identifier:
python data/placenames/placename2region.py data/placenames/placename_regions.csv data/trees/aebs.csv data/placenames/rin_region.csv
Now we have the files we will need for the analysis:
- The genealogy of the FarGen cohort:
data/trees/fargen.csv - The regional stratification of the individuals in the entire genealogy registry:
data/placenames/rin_region.csv - The FarGen IDs (FN) of the the individuals in the cohort:
data/ids/fargen_rin_samplename.csv