The representation learning models constructed in this research is used to describe the five biological features of proteins. These vectors together constitute the protein deep profile matrix.
Additional file 5: Table S2. The list of protein features utilized in the prediction model and the statistical significance differentiating the protein properties of drug targets from non-drug targets
The “Votes” column indicates how many feature selection algorithms cast a vote for that particular feature during the 10-fold cross-validation selecting procedure; the “Avg Rank” column describes the
For every feature, proteins are divided into a definite number of bins in terms of the different feature values. Then we use number(from 1 to 113) to represent different bins for different features. F
