K=2 is the best? Or K =9?

rajanil / fastStructure

A variational framework for inferring population structure from SNP genotype data.

MIT License

134 stars 50 forks source link

K=2 is the best? Or K =9? #55

Open pwang16 opened 4 years ago

pwang16 commented 4 years ago

I got this results below using chooseK.py from K=2 to 9:

Model complexity that maximizes marginal likelihood = 9 Model components used to explain structure in data = 2

Does that mean k=2 is best to explain most of the structure underlying the data? Or 9?

Thanks, Peng

liulu0827 commented 3 years ago

Hi, Peng. Do you get the answer? I have the same question.

Liu

pwang16 commented 3 years ago

If you can not find it in the stream, it is probably not answered.

On Fri, 9 Apr 2021 at 00:26, LuLiu @.***> wrote:

Hi, Peng. Do you get the answer? I have the same question.

Liu

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/rajanil/fastStructure/issues/55#issuecomment-816475116, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADXTZ3H3LRYMRLU7OFANA3LTH2TY3ANCNFSM4ODIHL7A .