Overlaying local ancestry on IBD segments

This pipeline is a series of three Python scripts that determine if both haplotypes in a long IBD segment have the same local ancestry. The final files are of the type chr.lai.ibd.overlap.gz under the folder you specify in the config.yaml file.

• The ANCESTRY column is a number that corresponds to those panels in the FLARE anc.vcf.gz output.
• The ANCESTRY is -1 if one or both of the haplotypes have an ancestry switch in the IBD stretch.
• The file format mirrors that of hap-ibd.jar but with an ANCESTRY column and a header.

In config.yaml, you need to adjust the file paths to where your genetic map, FLARE, and IBD segment data are.

You can run the script with snakemake -c1 --configfile config.yaml. Use -n to make a dry-run and see if the files are specified correctly. Visit isweep for more advice.

I have not designed the scripts/overlap-flare-lai-and-ibd.py script for chunked dataframe processing. If you have a large IBD file (say from a biobank-scale) analysis, the program may crash.

Assumptions

---

Some assumptions of the scripts are:

• There are less than 11 ancestry groups in the FLARE analysis.
• The FLARE optional parameter is probs=false (hard calls not ancestry dosages).
• You have used the file format of hap-ibd.jar to call IBD.
• You have Python pandas installed in your current environment.
• You have Snakemake installed in your current environment.

Other software

---

This code is part of a Snakemake workflow.

For more information about Snakemake workflows, rules, and configuration, see the official documentation: https://snakemake.readthedocs.io/.

This code is written and tested on output data from

• FLARE (Browning Lab)
• hap-ibd (Browning Lab)

This code can be run after the isweep local ancestry pipeline.