Starting at the variant of interest, the region is extended out ±0.1cM

Question

[tonja.r]

I might misunderstanding the following idea:

Starting at the variant of interest, the region is extended out ±0.1cM, using HapMap fine-scale estimates of recombination rates.

From the HapMap you get file containing the following:

chrom start    stop       Rate_cM.Mb             Avg_cM                          Gen_map_cM
Chr1 45413 72433     2.27700294070856   0.061526896460886      0
Chr1 72434 78031     2.42987467531016   0.0136024384323863    0.061526896460886
Chr1 78032 227743   2.43386172133186   0.364378306024035      0.0751293348932723

If a variant falls into 45413 - 72433 range then in order to calculate +- 0.1 cM region around it, I need to proceed as follows:

72411-45413 = 27009 bp , this region has an average 0.0615 cM, meaning that 0.1cM corresponds to 43917 bp ( 43917 = 27009*0.1/0.0615). Is it right?

Thank you in advance.

Answer 1

[Devon Ryan]

Not quite. Remember that the recombination rate changes with position, so the expected recombination rate over the next 43917 bases isn't going to be 0.1. So if we're interested in +0.1cM from 72411, then getting to 72433 gets us 4.844534e-05cM. Getting to 78031 adds an additional 1.36e-2, bringing us to 0.01365088. The means we need another 0.08634912cM. The last entry has an average rage of 2.434e-6cM/base, so dividing gives us an additional ~35478 bases. That puts us at 113509, meaning that the total width is 41098 bases.

The distance also won't be symmetric (i.e., -0.1cM probably won't be 41098 bases in the other direction).

[tonja.r]

I see where my mistake was. Thank you. The only small question I have is: How did you get 4.844534e-05cM?

I get the following: 72433-72411=22, 22*2.277e-6 = 5.0094e-05. Is it a rounding error or I am mixing something?

[Devon Ryan]

Good catch! It looks like I made a silly mistake and mistyped something into R. The exact numbers in my reply are a bit off, then, but the general concept is correct. I knew I should have had a coffee before posting...