search

StoreyLab / terastructure

TeraStructure is a new algorithm to fit Bayesian models of genetic variation in human populations on tera-sample-sized data sets (10^12 observed genotypes, i.e., 1M individuals at 1M SNPs). This package provides a scalable, multi-threaded C++ implementation that can be run on a single computer.
GNU General Public License v3.0
48 stars 9 forks source link

improved documentation #3

Closed whao89 closed 8 years ago

whao89 commented 8 years ago

especially with regards to --compute-betas

premgopalan commented 8 years ago

yes, please remove. also we need to remove variants of the algorithm A, B, C etc. -- i forget -- that are never used.

whao89 commented 8 years ago

good point, i'll add that as a issue since it's kind of separate.

Aquarius2016 commented 8 years ago

how does teraStructure generate a graph of distribution for theta? By the way, how can I find the max likelihood score among the initialization from k=3 to N?

whao89 commented 8 years ago

we generated the theta plots just drawing stacked bar plots in R (with no space between the bars). The validation likelihoods are in the file validation.txt associated with your run's output. It's the third column. You'll want to compare the last of those values for the different K you're interested in, ideally with a few reps.

whao89 commented 8 years ago

added a wiki