Hello,
I am trying to use the C software for estimating the iHS statistic (from Pritchard group http://hgdp.uchicago.edu/Software/). I am only looking at chromosome 12 in different populations from the 1000 genomes data and I have created the necessary files with the haplotypes and the legends (using only SNPs with MAF of over 5%), but when I run it, I only get the score for about 160 SNPs from the whole chromosome. Has anyone had this problem before? Would anyone know if I need to split the chromosome in windows of numbers of SNPs or of physical position?
The iHS script is slow and sloppy, and sometimes stops with a segmentation error and no documentation. You can try this other version, from the same author: iHS from WHAMM!
In any case, the iHS binary from the Pickrell page should not stop after only 160 SNPs. A possible source of error can be that you are using a recombination map instead of a genetic map. This is the most common error when using the iHS script.
Genetic maps are the integral of the recombination maps. If you are not sure about whether you are using a recombination map or a genetic map, you can check two criterias:
values in a genetic map should range from 0 to 1. If any of the values in your file is higher than 1, then you are using a recombination map.
in a recombination map, each row contains the recombination frequency relative to the previous SNP in the file. So, in the example below, the value of 0.5 at position 30000 means that the recombination frequency between 15000 and 30000 is 0.5. On the contrary, in a genetic map, all the scores are relative to the first position in the chromosome. So, each row should be larger than the previous.
Check that you are using a genetic map, and try running iHS again.
Position Recombination map Genetic map
1 0 0 (note: numbers are made up, they don't correspond between the two columns)
300 0.3 0.03
600 0.2 0.05
1000 1.4 0.14
15000 0.2 0.16
30000 0.5 0.21
Thank you for the answer, it sounds like my problem may be in the genetic positions but I still don't fully understand what I've done wrong. As you guessed, my positions do not range from 0 to 1, but the positions are all relative to the 1st SNP of the file, since they are incrementing along with the physical position. I got the files from the Rutgers website (http://compgen.rutgers.edu/mapinterpolator/) which is supposed to be giving the (sex specific) genetic positions (but used the average cM). Where would you recommend I get the genetic positions from?
In addition, looking at the sample files provided with the software, the column that is for the genetic positions goes over 1 as well, so I don't understand how that can be..?
Hi,
I have done some analyses about positive selection and I have calculated iHS.
I used as genetic maps those from hapmap http://hapmap.ncbi.nlm.nih.gov/downloads/recombination/2008-03_rel22_B36/.
But the positions in these map don't range range from 0 to 1. I think I'm missing something...can you help me?
Thank you
Thank you for the answer, it sounds like my problem may be in the genetic positions but I still don't fully understand what I've done wrong. As you guessed, my positions do not range from 0 to 1, but the positions are all relative to the 1st SNP of the file, since they are incrementing along with the physical position. I got the files from the Rutgers website (http://compgen.rutgers.edu/mapinterpolator/) which is supposed to be giving the (sex specific) genetic positions (but used the average cM). Where would you recommend I get the genetic positions from? In addition, looking at the sample files provided with the software, the column that is for the genetic positions goes over 1 as well, so I don't understand how that can be..?
Hi, I have done some analyses about positive selection and I have calculated iHS. I used as genetic maps those from hapmap http://hapmap.ncbi.nlm.nih.gov/downloads/recombination/2008-03_rel22_B36/. But the positions in these map don't range range from 0 to 1. I think I'm missing something...can you help me? Thank you