Skip to main content

Table 3 Summary of different features for the inhibitors used in our study

From: Predicting a small molecule-kinase interaction map: A machine learning approach

Short-hand Full Name # Features Feature Type
PT Primary Target 1 nominal
Determine for which kinase(s) an inhibitor shows a preferred binding
MS 2D Molecular Structure 1 nominal
Clustering of inhibitors due to their 2D structure
Cluster1 = {SB202190, SB203580}
Cluster2 = {CI-1033, EKB-569, MLN-518}
Cluster3 = {Staurosporine; LY-333531}
Cluster4 = {Roscovitine; Flavopiridol; BIRB-796}
Cluster5 = {SU 11248; BAY-43-9006; ZD-6474; Gleevec,
GW-2016, Iressa, Tarceva, V X-745, V atalanib, SP600125}
FTs Free Trees 78 numeric
Determine frequently occurring acyclic substructures as structural features
KNN KNN clustering 20 numeric
Detect each inhibitor's k nearest neighbors
CF Chemical Features 15 numeric
Calculate various chemical features with JOELib2
GF Geometric Features 5 numeric
Calculate various geometric features
P Pharmacophores 50 numeric
Calculate 3-point pharmacophores