IMPUTE2- Map file for custom reference panel

Question

[john.michel.rouhana]

Hi all,

I'm trying to build a custom reference panel for use with IMPUTE2. I have a phased .vcf from all my participants, and I was using the first script here, modified slightly to work with my vcf/HG38, to build the reference panel. The exact command I run to generate the files for chromosome 1 (as an example) is:

./vcf2impute_legend_haps.pl -vcf my_vcf.vcf.gz \
  -leghap output_prefix \
  -chr chr1  \ 
  -hg38

The -hg38 thing is just an argument I added to the script, and can be ignored.

This generates .hap.gz files, .legend.gz files, and .sample_list files, one for each chromosome. What I'm missing is a genetic map file. I don't what these are, and I can't find anything on it in the documentation. Looking at the 1000 Genomes HG37 reference panel files provided here, it's not clear to me what these mapping files are, so I can't replicate the steps to create them.

Any advice would be appreciated.

[rdmorin]

Did you ever get impute2 working with hg38 coordinates?

[Jacky]

https://alkesgroup.broadinstitute.org/Eagle/downloads/tables/ I found this genetic map file that contains all autosomes and Chr X in HG38 coordinate.

[Paravee]

Hi. The genetic map you posted is provided by Eagle, but it can be use with IMPUTE2 too right? I'm rather new to imputation, so I want make sure. Thanks.

Answer 1

[h.mon]

According to the documentation you linked, genetic maps are recombination maps, measured in centiMorgans.

If you can lift-over the genomic position of the markers from the 1000 Genomes HG37 reference panel, you can use the same genetic map position (the Genetic_Map(cM) column), and recalculate the recombination rate between current position and next position in map (the COMBINED_rate(cM/Mb) column).

However, are you sure there isn't an updated version of these files?

[john.michel.rouhana]

I tried to find updated versions. What you say makes sense, and is actually the conclusion I came to (I missed the response). Once I revisit this problem I'll likely be resorting to recalculating the recombination rates if I cannot find updated files.