search

dougspeed / LDAK

Other
12 stars 1 forks source link

Can LDAK calculate reproductive traits? #6

Open huangxiaoyu0 opened 1 month ago

huangxiaoyu0 commented 1 month ago

Hi dougspeed, Recently, I was trying to estimate the heritability of reproductive traits using LDAK, but the error showed multiple individual numbers, wondering whether LDAK can not directly estimate reproductive traits.

Best regards, Huang xiaoyu

dougspeed commented 1 month ago

Hi Huang, thanks for the question

It sounds like your phenotype has multiple values for some / all individuals?? Sorry, LDAK can not handle these types of phenotypes (there can only be at most one value per individual).

You are welcome to give more details of your phenotype, and desired analysis, and I can give advice / see if it is something I can accommodate in LDA

Thanks Doug

huangxiaoyu0 commented 1 month ago

Hi Doug, thanks for getting back to me so quickly about this.

My analytical focus lies in the total litter performance of sows, where each sow exhibits varying total litter sizes across different parities. I aim to apply the LDAK software to calculate the repeatability model, with the objective of comparing its outcomes against those obtained using DMU , to ascertain whether similar conclusions can be drawn.Because I want to evaluate genomic selection accuracy with LDAK.

Best regards, Huang xiaoyu

dougspeed commented 1 month ago

Thanks for the details. I am not familial with repeatability models, but it seems they look like Model 1 in this paper https://www.sciencedirect.com/science/article/pii/S002203021300060X, where there are multiple random effects, nested within each other. I think this type of model could in theory be fit within LDAK, by using the REML function and specific kinship matrices, but would not be straightforward.

I will think more about it. If you have a reference that provides a mathematical model, that would be helpful Thanks

JasperHof commented 1 month ago

Hi Huang,

I believe LDAK currently does not include any models that are targeted at repeated measures. Although these could potentially be analysed using mixed models (such as LDAK-KVIK https://www.ldak-kvik.com/), its application on repeated data has not been tested.

However, your data might be suitable for analysis in TrajGWAS? https://www.cell.com/ajhg/fulltext/S0002-9297(22)00049-0 The focus for this method lies in different values over time, but it might be applicable to your situation.

Best,

Jasper

huangxiaoyu0 commented 1 month ago

Hi Jasper,

Thank you for your suggestion, I will try it out when I have time.

Best, Huang xiaoyu