Fig. 2
From: SeqScreen: accurate and sensitive functional screening of pathogenic sequences via ensemble learning
SeqScreen overview. A SeqScreen Workflow: this figure outlines the various modules and workflows of the SeqScreen pipeline. Boxes in green indicate that these modules are only run in the sensitive mode. The boxes in yellow are run in the fast mode, while the ones in blue are common to both modes. In addition to the two different modes, SeqScreen also contains optional modules that can be run based on the parameters provided by the user. B SeqScreen Human-in-the-loop Framework: includes initial annotation and curation of training data by manual curation. The data is used to train Ensemble ML models. The results obtained and selected feature weights are passed on back to biocurators to fine tune features and UniProt queries which form a new set of refined training data for the Ensemble model