Biological Image Indexing for Content-Based Retrieval of Drug Effects in Phenotypic Screening Data of Macroparasites

Abstract

Phenotypic-screening involves systematically assessing the therapeutic effects of a set of molecules by exposing entire disease systems to them and observing, through imaging, the effects of the compounds. Phenotypic assays typically generate hundreds of thousands to millions of images. An unmet challenge in this setting is to identify similar phenotypic effects caused by molecules, which may potentially be structurally different. While phenotypes can be compared using their feature vectors, real-time querying of these data sets becomes a challenging task because of the size of the data sets and the high dimensionality of the feature vectors. In this paper, we present an indexing approach that seeks to address this problem and allows efficient query-retrieval of phenotypic drug effects.