These are chat archives for theboocock/selectionTools
I have had this problem in my own data and did change the way multicore_iHH.R rejoined the individual iHH files after the calculation step.
The new method was updated into the selectionTools1.1 branch. It requires running the also updated haps_interpolate step before the updated multicore_iHH.R
Changing window size is valid, ideally a larger window size is used where possible so that the ehh decay for the SNPs isn't prematurely stopped at the window boundary
[.data.frame(map_positions, , 2) : undefined columns selected
*tmp*, , 3, value = 16) :