Question about ancestral/derived allele.

[song88180]

Hello,

I have a few questions about calculating iHS:

1. The original iHS paper (https://doi.org/10.1371/journal.pbio.0040072) calculated raw iHS by ln(iHH_ancestral/iHH_derived), but here, iHS is calculated by ln(iHH_derived/iHH_ancestral), so that positive iHS value means positive selection on the derived allele. Can you confirm if my understanding is correct?

2. Is it necessary to know which allele is the ancestral allele and which is the derived allele to calculate the iHS? For example, a SNP has two alleles A and B, but I don't know which one is the ancestral/derived allele. If I calculate raw iHS by ln(iHH_A/iHH_B) and the standardized iHS is positive, can I just say A is under positive selection?

3. Thank you so much for sharing your pre-calculated iHS here (https://datashare.ed.ac.uk/handle/10283/714). I wonder when you calculate those iHS, did you decide which allele is ancestral allele and which is derived allele first, and then always do ln(iHH_derived/iHH_ancestral) when calculating iHS? Or did you just use the reference/alternative coding provided by the 1000 Genome Project, and do ln(iHH_alternative/iHH_reference) ?

Thanks a lot!
Siliang

[prenderj]

Hi Siliang Regarding your points: 1) Yes the ratio is the other way round from the original iHS paper. 2) Strictly speaking iHS is meant to be calculated with respect to ancestral and derived alleles. Though people do sometimes just calculate as reference and alternate (and potentially take the absolute of the returned value). Just need to be careful with interpretation. If you dont have an ancestral genome for your species of interest you can though actually generate it using one of our other tools ( https://github.com/evotools/nSPECTRa). This takes as input a cactus multi-species alignment file (for example for mammals can take the zoonomia HAL file https://zoonomiaproject.org/the-data/). Then you can polarise your alleles to ancestral and derived using this. Although the rest of the workflow generates mutation spectra you can stop once get the ancestral genome. 3) Apologies, I cant actually remember if we polarised by ancestral allele this data. But would be relatively quick to rerun a chromosome and check. Cheers James

…

On Sun, 1 Jun 2025 at 07:14, SSL ***@***.***> wrote: *song88180* created an issue (evotools/hapbin#69) <#69> Hello, I have a few questions about calculating iHS: 1. The original iHS paper (https://doi.org/10.1371/journal.pbio.0040072) calculated raw iHS by ln(iHH_ancestral/iHH_derived), but here, iHS is calculated by ln(iHH_derived/iHH_ancestral), so that positive iHS value means positive selection on the derived allele. Can you confirm if my understanding is correct? 2. Is it necessary to know which allele is the ancestral allele and which is the derived allele to calculate the iHS? For example, a SNP has two alleles A and B, but I don't know which one is the ancestral/derived allele. If I calculate raw iHS by ln(iHH_A/iHH_B) and the standardized iHS is positive, can I just say A is under positive selection? 3. Thank you so much for sharing your pre-calculated iHS here ( https://datashare.ed.ac.uk/handle/10283/714). I wonder when you calculate those iHS, did you decide which allele is ancestral allele and which is derived allele first, and then always do ln(iHH_derived/iHH_ancestral) when calculating iHS? Or did you just use the reference/alternative coding provided by the 1000 Genome Project, and do ln(iHH_alternative/iHH_reference) ? Thanks a lot! Siliang — Reply to this email directly, view it on GitHub <#69>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AB5OYGX7ZCZRAHTNWXCRESL3BKK57AVCNFSM6AAAAAB6KXAX7WVHI2DSMVQWIX3LMV43ASLTON2WKOZTGEYDMNRZGM4DGOA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[song88180]

Hi James,

Thank you very much for your explanation!

Please let me know if you do get a chance to check if the shared data was polarized by ancestral allele or not, that would be really helpful!

Best!
Siliang