search

kgori / sigfit

Flexible Bayesian inference of mutational signatures
GNU General Public License v3.0
33 stars 8 forks source link

Replace sigfit.stan with sigfit_multi.stan #1

Closed baezortega closed 7 years ago

baezortega commented 7 years ago

Since almost any user will have multiple samples, and since the fitting model for one sample is just a special case of the multiple-sample one, I would suggest removing the sigfit.stan model and using sigfit_multi.stan always (perhaps renaming it to "sigfit_fitting.stan"), as well as removing the "multi" argument in the fit_signatures() function. That would also make the structure of the stanfit object coming from signature fitting consistent regardless of number of samples (always tri-dimensional).

baezortega commented 7 years ago

Implemented, but needs testing! The new model for fit_signatures() – with hierchical=FALSE – which is the previoius sigfit_multi.stan, is now called sigfit_fit.stan.