Fig. 1.

Schematic of the selection process performed by submodular selection of assays (SSA). The method takes as input all available existing genomics assays, where each assay is represented as a real-valued track over the genome. In the SSA-past mode, SSA selects a panel of already-performed assays to use as input to an expensive computational analysis. In the SSA-future mode, SSA chooses a panel of assay types to be performed in a new cell type. In both cases, the resulting data sets are provided as input to downstream analyses, which may include imputing assays that were not performed, predicting the locations of functional elements, or semi-automated genome annotation. SSA submodular selection of assays