search

jbloomlab / phydms

phylogenetic analyses informed by deep mutational scanning data
GNU General Public License v3.0
14 stars 4 forks source link

Q values should be for one sided tests? #15

Closed jbloom closed 6 years ago

jbloom commented 7 years ago

For the diversifying selection tests, it seems like perhaps the Q values should be computed independently / differently for omega > 1 versus omega < 1.

jbloom commented 7 years ago

This has been fixed in version 2.0.0.

Note that this will change the omega values in the omegabysite.txt files.

@skhilton: Can you update the tutorial to show the new values that will be computed? They are now going to be slightly different than in the current tutorial. Once that is fixed, you can close this issue.

@Haddox: Note that this will slightly change the sets of values that are significantly different from natural evolution. This should resolve the problem that Alexandria noted, where sometime there were more sites of diversifying selection values identified for YNGKP than for ExpCM. That is because the negative selection sites were inflating the Q values of the diversifying selection values.