Reasoning with chemically induced dynamic phenotypes in whole-organism assays

点击翻译按钮获取中文摘要

ABSTRACT Schistosomiasis, caused by a parasitic flatworm, is a disease of the poor. It infects over 200 million people worldwide and places another 800 million at risk. Current treatment and control of this disease relies on just one drug, praziquantel (PZQ) - a precarious situation should drug resistance emerge. The therapeutic profile of PZQ is also not ideal. The World Health Organization has therefore declared schistosomiasis a disease for which new therapies are urgently needed. Drug discovery for schistosomiasis in particular (and helmintic diseases in general) is traditionally based on phenotypic screening, whereby the parasite(s) are exposed to compounds and their systemic responses, such as changes in shape, appearance, and motion, are analyzed to identify hits. Analyzing the temporally varying, high-dimensional, and information-rich output from phenotypic screening of a complex macroparasite is, however, non-trivial. This fact is underlined by the complete absence of any database(s) or analysis tools for disease-causing helminths that would allow analysis and reasoning with dynamic phenotypic data. To address this need, we formulate the following two aims: Under Aim 1, we propose to develop the first quantitative and publicly available database of the schistosome’s time- varying response to chemical probes. The database will support content-based querying of dynamic phenotypes using time-series matching. The information in this database will underpin structure-activity relationship (SAR) studies with the drug targets and associated small molecule chemistries that we have validated. This phenotypic record will also aid understanding of the molecular mechanism of action (MMoA) of various chemistries and serve as a reference for phenotypes elicited using other compounds by researchers worldwide. Under Aim 2, we will develop algorithmic methods for analyzing the time-varying phenotypic responses of the schistosome parasite. These methods will allow scientists to match, compare, cluster, and quantitatively reason-with dynamic (i.e. temporally varying) phenotypes. In particular, scientists will be able to: (1) objectively compare phenotypic responses of parasites to identify similar effects, even when they occur due to structurally distinct compounds, (2) relate phenotypic effects observed in different studies conducted under varying conditions, (3) stratify the phenotypic variability within and across parasite populations, and (4) prioritize compounds based on quantitative reasoning with dynamic and complex phenotypic responses. Results from both aims will be made freely available to biologists worldwide through a public database and software developed by us. Our proposal constitutes an innovative point of progress in (a) developing algorithmic methods and datasets for reasoning-with and understanding the phenome of the etiological agent of schistosomiasis and leveraging it for drug discovery and (b) establishing a rigorous analysis framework and publicly available resources that can be applied to other complex disease-causing macroparasites.