major difference in number of pairwise ibd events between chromosome 1 and other autosomes

[cteerlink101]

Hi,

I am doing an analysis where I have noticed that hap-ibd results in a much larger number of pairwise IBD events on chromosome 1 compared to the other chromosomes. Here are a few file sizes (in bites) to illustrate what I am seeing:

VCF (input file)
chr1: 1.577 Gb
chr2: 1.559 Gb
chr3: 1.314 Gb

IBD (output of hap-ibd)
chr1: 44.475 Gb
chr2: 2.301 Gb
chr3: 1.003 Gb

You can see that chr1 result file is 20X the size of other chromosomes, although the input files (vcf's) are of similar scale.

Have you come across this before, or do you have any ideas what might be producing this anomaly?

Also, I am using the b38 genetic maps available from the hap-ibd github, and I am using v1.0 of hap-ibd (I do not know if the version available from your github site is more current than that).

Any help is appreciated.

Thanks!

[browning-lab]

Regions with low marker density or without markers can cause false positive IBD segments when IBS segment length is used to detect IBD segments. For example, a 2 cM IBS segment in an outbred human population is unlikely to be an IBD segment if it contains a 1.5 cM interval that has no markers. You can test whether this is the case by using a grid of points across the chromosome (e.g. with a 100-kb spacing) and plotting the number of output IBD segments covering each grid point on the chromosome. IBD coverage should be relatively uniform, except near chromosome ends. If you see a large spike in IBD coverage, look for gaps in marker coverage under the spike.