morrislab / metient Goto Github PK

Hi,

Thanks for making this tool available the pre-print was fantastic.

I wanted to ask if you had any experience running metient in a situation where the primary site is unknown. Say I have ten samples where I am unsure which is the primary site - should I input "primary" for all samples in the site_category table? My understanding is that it would run metient ten times, each time using a different sample as a different "true primary." When metient runs (I am planning on using the evaluate mode since I have a low case number) will it compare the graph scores only within a primary designation (and output the most likely graph for each primary designation, but not comparison between primary designations), or will it compare the scores of the graphs also between different primary designations.

I see in supplementary figure S3 for patient 1 it output three solutions, with the graph on the fair right having the primary designation "left ovary" instead of right ovary as the others do. How was metient run in this case i.e. did you input for both left and right ovary primary in the site_category_table? This case is particularly interesting because in the machina paper they argue given migration/comigration numbers the tumor likely originated from the left ovary (although room for interpretation).

Thank you again looks like a great tool!