Given that gene-centric (GBP) models are trained separately for each GO term, it stands to reason that performance must degrade for GO terms with very few annotated genes to use as positive training examples.
For GO terms in evaluation categories with association counts in the ranges [11,30], [31,100], and [101,300], the α coefficients are determined separately for each GO term.
Epochs were averaged separately for each condition.
Significance was assessed separately for each cross-product term (Wald test), as well as overall for the multivariable model with both cross-product terms versus neither (likelihood ratio test).
Separately for each GO-process term ti, three measurements, precision pi, recall ri and F-measurement fi were computed.
Checking that the map indeed coincides with (8.42) can be done separately for each of the terms ({text {Hom}} overline{L}_i(A_0),overline{M}_i(B_0))), (i in I) of (8.27).
O-H tests were performed separately for each species, if the interaction term sex*patch reached significance we computed the tests separatedly for males and females, otherwise the sexes were pooled.
Models are presented separately for each cohort, and include an interaction term for maternal and paternal mental health.
To explore possible interactions we fitted interaction terms (model 2) separately for each drug or drug class by age (>65 or ≤65), sex, aspirin use, and statin or fibrate use.
Therefore, we removed long-term trends in ePAD separately for each person via spline-based nonparametric regression.
