Recommended length of burnin period and number of mcmc reps after burnin

[Tobias Kox]

Hi,
I am using a dataset consisting of over 15000 SNPs and over 1000 different genotypes(lines). Which numbers are recommended to use for length of burnin period and mcmc reps. To have a good balance on accuracy and computation time. I am using a newer computer i7 and 16GB ram.

Second smaller question:
Is there anay limitation in structure for size of dataset in regard to number of marker or number of genotypes?

Hope to get some good answers.

Thanks a lot.

Tobi

[Vikram Chhatre]

Structure imposes no limit on the size of the data set but it does get computationally very intensive once you venture into larger data sets.  I recommend looking at fastSTRUCTURE, which can gracefully handle genome wide data sets.

V

--
You received this message because you are subscribed to the Google Groups "structure-software" group.
To unsubscribe from this group and stop receiving emails from it, send an email to structure-softw...@googlegroups.com.
To post to this group, send email to structure...@googlegroups.com.
Visit this group at http://groups.google.com/group/structure-software.
For more options, visit https://groups.google.com/d/optout.

[Vikram Chhatre]

Sorry forgot to answer the other question.  You need as many burnin reps as it will take for your data to converge.  Generally 10,000 burnin is enough, but at least 50,000 is highly recommended.  Collect data for longer.

V

[Ahmed Sallam]

Hi Vikram

The results of fastSTRUCTURE are quite different from the results obtained by STRUCTURE, which always agreed with PCA.
https://www.nature.com/articles/s41437-019-0247-6

egaer to receive your reply