Skip to main content
. 2019 Jun 4;36(1):250–256. doi: 10.1093/bioinformatics/btz470

Fig. 1.

Fig. 1.

Template’s general workflow with TPOT-FSS and example pipelines. Final pipelines with optimized parameters are shown for simulated data (top, green) and real-world RNA-Seq data (bottom, mauve). The specific operators selected in optimal pipelines include built-in TPOT’s operators (OneHotEncoder, FeatureSetSelector) and functions from the library of scikit-learn (ExtraTreesClassifier, Nystroem)