Learning perturbation-inducible cell states from observability analysis of transcriptome dynamics

Abstract

A major challenge in biotechnology and biomanufacturing is the identification of a set of biomarkers for perturbations and metabolites of interest. Here, we develop a data-driven, transcriptome-wide approach to rank perturbation-inducible genes from time-series RNA sequencing data for the discovery of analyte-responsive promoters. This provides a set of biomarkers that act as a proxy for the transcriptional state referred to as cell state. We construct low-dimensional models of gene expression dynamics and rank genes by their ability to capture the perturbation-specific cell state using a novel observability analysis. Using this ranking, we extract 15 analyte-responsive promoters for the organophosphate malathion in the underutilized host organism Pseudomonas fluorescens SBW25. We develop synthetic genetic reporters from each analyte-responsive promoter and characterize their response to malathion. Furthermore, we enhance malathion reporting through the aggregation of the response of individual reporters with a synthetic consortium approach, and we exemplify the library’s ability to be useful outside the lab by detecting malathion in the environment. The engineered host cell, a living malathion sensor, can be optimized for use in environmental diagnostics while the developed machine learning tool can be applied to discover perturbation-inducible gene expression systems in the compendium of host organisms.

A Data-Driven Koopman Approach for Power System Nonlinear Dynamic Observability Analysis

“…Now, we would like to mention that the reason we select the EDMD method instead of the aforementioned DMD method or the modified Arnoldi method [32], [33], [39], [46], is because the key step to formulate the proposed augmented Koopman operator relies on (22). Therefore, the DMD method is not appropriate for the formulation adopted in this article.…”

Accelerating Genetic Sensor Development, Scale-up, and Deployment Using Synthetic Biology

“…By monitoring global gene expression in response to a target, it is possible to build a predictive model by correlating the input of the target to changes in the activity of hundreds or even thousands of promoters. Advancements in NGS, microfluidics, and machine learning have now made this form of differential sensing possible [ 138 , 139 ].…”

“…In a complementary study, transcriptomics was directly used to characterize the promoter activity of approximately 6,000 genes in Pseudomonas fluorescens in response to exposure to the pesticide malathion [ 139 ]. Although transcriptomics can be expensive, the cost of sequencing has dramatically dropped over the past decade, making this approach feasible for even small laboratories.…”

“…The use of NGS to uncover innate responses of microbes may offer a route forward. NGS techniques are beginning to be used to identify native promoters within diverse microbial hosts that respond to a target of interest [ 137 , 139 , 217 ]. Given the stringent regulations surrounding genetically modified organisms for real-world use, it would be advantageous to leverage the inherent capabilities of organisms already present where a system needs to be deployed.…”

Data-driven learning of structure augments quantitative prediction of biological responses

“…Biological systems respond to a wide range of environmental cues, including temperature [1][2][3], nutrient variations [4][5][6], and stresses [7][8][9]. On one hand, these responses reflect the intrinsic properties of each system; on the other hand, they serve as a tunable property that is being widely exploited [10][11][12]. This type of manipulation has allowed us to maximize bacterial growth for synthesis of desirable chemicals, such as bioplastic [13][14][15], biofuels [16][17][18] and artificial flavors in beer [19,20].…”