Representing high throughput expression profiles via perturbation barcodes reveals compound targets.

High throughput mRNA expression profiling can be used to characterize the response of cell culture models to perturbations such as pharmacologic modulators and genetic perturbations. As profiling campaigns expand in scope, it is important to homogenize, summarize, and analyze the resulting data in a manner that captures significant biological signals in spite of various noise sources such as batch effects and stochastic variation. We used the L1000 platform for large-scale profiling of 978 representative genes across thousands of compound treatments. Here, a method is described that uses deep learning techniques to convert the expression changes of the landmark genes into a perturbation barcode that reveals important features of the underlying data, performing better than the raw data in revealing important biological insights. The barcode captures compound structure and target information, and predicts a compound's high throughput screening promiscuity, to a higher degree than the original data measurements, indicating that the approach uncovers underlying factors of the expression data that are otherwise entangled or masked by noise. Furthermore, we demonstrate that visualizations derived from the perturbation barcode can be used to more sensitively assign functions to unknown compounds through a guilt-by-association approach, which we use to predict and experimentally validate the activity of compounds on the MAPK pathway. The demonstrated application of deep metric learning to large-scale chemical genetics projects highlights the utility of this and related approaches to the extraction of insights and testable hypotheses from big, sometimes noisy data.

Parsimonious discovery of synergistic drug combinations.

Combination therapies that enhance efficacy or permit reduced dosages to be administered have seen great success in a variety of therapeutic applications. More fundamentally, the discovery of epistatic pathway interactions not only informs pharmacologic intervention but can be used to better understand the underlying biological system. There is, however, no systematic and efficient method to identify interacting activities as candidates for combination therapy and, in particular, to identify those with synergistic activities. We devised a pooled, self-deconvoluting screening paradigm for the efficient comprehensive interrogation of all pairs of compounds in 1000-compound libraries. We demonstrate the power of the method to recover established synergistic interactions between compounds. We then applied this approach to a cell-based screen for anti-inflammatory activities using an assay for lipopolysaccharide/interferon-induced acute phase response of a monocytic cell line. The described method, which is >20 times as efficient as a naïve approach, was used to test all pairs of 1027 bioactive compounds for interleukin-6 suppression, yielding 11 pairs of compounds that show synergy. These 11 pairs all represent the same two activities: ß-adrenergic receptor agonists and phosphodiesterase-4 inhibitors. These activities both act through cyclic AMP elevation and are known to be anti-inflammatory alone and to synergize in combination. Thus we show proof of concept for a robust, efficient technique for the identification of synergistic combinations. Such a tool can enable qualitatively new scales of pharmacological research and chemical genetics.

The discovery of a potent and selective lethal factor inhibitor for adjunct therapy of anthrax infection.

A potent and selective anthrax LF inhibitor 40, (2R)-2-[(4-fluoro-3-methylphenyl)sulfonylamino]-N-hydroxy-2-(tetrahydro-2H-pyran-4-yl)acetamide, was identified through SAR study of a high throughput screen lead. It has an IC50 of 54 nM in the enzyme assay and an IC50 of 210 nM in the macrophage cytotoxicity assay. Compound 40 is also effective in vivo in several animal model studies.