Prevalence and patterns of higher-order drug interactions in Escherichia coli

Tekin, Elif; White, Cynthia; Kang, Tina Manzhu; Singh, Nina; Cruz-Loya, Mauricio; Damoiseaux, Robert; Savage, Van M.; Yeh, Pamela J.

doi:10.1038/s41540-018-0069-9

Cited by 75 publications

(101 citation statements)

References 40 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To investigate how species interactions changed with diversity, we characterised the species interactions of the communities using the long-term coexistence experiments. We modified a multiplicative framework originally developed to help quantify higher order antibiotic interactions on Escherichia coli (Tekin et al . 2018a).…”

Section: Methodsmentioning

confidence: 99%

Characterising a stable five-species microbial community for use in experimental evolution and ecology

Padfield

Castledine

Pennycook

et al. 2020

Preprint

View full text Add to dashboard Cite

23The ability of species to mutually invade from rare is the defining measure of species 24 coexistence. However, it is unknown whether invasion growth rates predict any characteristic 25 of long-term community dynamics. Here, we use a model five-species microbial community 26 to investigate the link between short-term growth rate and long-term relative abundance. We 27 manipulated diversity and tested the ability of species to coexist in different combinations. 28Across all diversity levels and species combinations, populations re-established from rare in 29 71 of 75 combinations and all combinations were stable in long-term culture. Moreover, short-30 term relative invader growth rate was positively associated with long-term equilibrium 31 proportion, despite large variation in interactions between species and communities. This 32 finding was confirmed using a modelling approach and suggests that the short-term invasion 33 growth rate can predict long-term relative abundance within that community. 34 56 both support (Martínez-Meyer et al. 2013) and refute (Dallas & Hastings 2018) the niche 57 centrality hypothesis across animal and plant communities. Recent theory has highlighted why 58 this macroecological relationship may not be ubiquitous, with spatial and temporal variation 59 3 and metapopulation dynamics potentially causing this relationship to break down (Holt 2019; 60

show abstract

Section: Methodsmentioning

confidence: 99%

Characterising a stable five-species microbial community for use in experimental evolution and ecology

Padfield

Castledine

Pennycook

et al. 2020

Preprint

View full text Add to dashboard Cite

show abstract

“…The prospects of higher order synergies ( i.e ., interactions beyond pairwise) and scaling laws for drug mixtures, while provocative, have remained contentious [29, 8, 30]. MuSyC’s cubic geometry allows it to be easily extended to three or more drugs (Figure S1), and we expect MuSyC will enable a more refined search for higher order interactions.…”

Section: Discussionmentioning

confidence: 99%

A Consensus Framework Unifies Multi-Drug Synergy Metrics

Wooten

Meyer

Quaranta

et al. 2019

Preprint

View full text Add to dashboard Cite

10Drug combination discovery depends on reliable synergy metrics; however, no consensus exists on the 11 appropriate synergy model to prioritize lead candidates. The fragmented state of the field confounds 12 analysis, reproducibility, and clinical translation of combinations. Here we present a mass-action based 13 efficacy masks synergistic interactions; (2) MSP frameworks are biased toward antagonism for drugs with 54 intermediate efficacy; and (3) DEP frameworks contain a Hill-slope dependent bias. The Hill-slope bias 55 results from satisfying the famous "sham" combination thought experiment, arguing against the merit of 56 sham-compliance as a measure of validity for synergy frameworks. Using two large combination datasets 57 [23, 24], MuSyC identifies real-world examples where the conflicting assumptions of previous drug synergy 58 frameworks misleads or impedes drug discovery efforts through these pervasive and predictable biases. We 59 therefore propose MuSyC as a consensus framework to interpret combination pharmacology and signify its 60 broad applicability to the study of drug mixtures. 61A multi-dimensional formalism to measure multi-drug synergistic effects 62 Low Efficacy High Efficacy High Cooperativity Low Cooperativity h Low Potency High Potency The 4-parameter Hill equation is commonly used to fit dose-response data from in vitro and in vivo assays 63 (see Box 1 equation (3) for derivation and Table 1 for parameter annotation). This equation can be derived64 from the equilibrium of a 2-state model of drug effect based on the Law of Mass Action (Figure 1A left).65 Traditionally, the parameters of the Hill equation are interpreted as a drug's efficacy (E 0 − E 1 ), potency 66 (C), and cooperativity (h), also known as the Hill slope. These parameters correspond to three possible 67 geometric transformations of a dose-response curve (Figure 1A right). To generalize this one-drug formalism 68 to two concurrent drugs, we recently developed a 4-state mass-action model of combination pharmacology 69 (Figure 1B left) [21]. From this model, we derived a two-dimensional (2D) Hill equation for two drugs 70 2 (Box 1 equation 8) defining a dose-response surface (Figure 1B middle). The 2D Hill equation contains five 71 additional parameters, not present in the single-drug Hill equation, which measure different types of drug 72interactions. These additional parameters measure changes in a drug's efficacy (β), potency (α 12 and α 21 ), 73 and cooperativity (γ 12 and γ 21 ) in a combination -representing three distinct types of synergy ( Figure 1B 74 right, Table 1). As we show below, these parameters are conflated in traditional synergy metrics obscuring 75 the true origin and magnitude of drug synergy or antagonism. 76 A bridge between DEP and MSP maps existing synergy approaches onto a com-77 mon landscape 78 In recent years, several alternative synergy models have been proposed. Broadly, these models are derived 79 from one of two guiding principles: the Multiplicative Survival Principle (MSP) and the Drug...

show abstract

“…Directing our attention to the second plot in Figure 2, recall that the maximizing distributions for Cohesion-2 over four binary variables was given in Table II. By the result of Matus, we know that over a sufficiently large field, D(p||p (2) ) and Cohesion-2 have the same global maximizers, but what about when the support is restricted? When optimizing over a discretized simplex of bin size 1/12, the maximizing distribution for D(p||p (2) ) over four binary variables is While this might not be the global maximizer, we can still gain insight from comparing it to maximizers of Cohesion-2.…”

Section: Related Workmentioning

confidence: 99%

“…By the result of Matus, we know that over a sufficiently large field, D(p||p (2) ) and Cohesion-2 have the same global maximizers, but what about when the support is restricted? When optimizing over a discretized simplex of bin size 1/12, the maximizing distribution for D(p||p (2) ) over four binary variables is While this might not be the global maximizer, we can still gain insight from comparing it to maximizers of Cohesion-2. This distribution does not have a linear dependence structure, since no marginal variables have entropy of 1 bit.…”

Section: Related Workmentioning

confidence: 99%

“…Characterizing the interactions in a system by its pairwise correlations is a common modeling assumption, but a more complete picture would also consider many-to-one, and many-to-many interactions. In systems with emergent properties, such as those studied in ecology [1], systems biology [2], neuroscience [3], and genetics [4] (to name a few), these higher-order interactions play a vital role in macroscopic behavior. The issue of defining a proper measure of higher-order correlation, at least from an information theoretic perspective, is what the study of multivariate mutual information (MMI) seeks to do.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Maximizing Multivariate Information With Error-Correcting Codes

Reing

Steeg

Galstyan

2020

IEEE Trans. Inform. Theory

View full text Add to dashboard Cite

Multivariate mutual information provides a conceptual framework for characterizing higher-order interactions in complex systems. Two well-known measures of multivariate information-total correlation and dual total correlationadmit a spectrum of measures with varying sensitivity to intermediate orders of dependence. Unfortunately, these intermediate measures have not received much attention due to their opaque representation of information. Here we draw on results from matroid theory to show that these measures are closely related to error-correcting codes. This connection allows us to derive the class of global maximizers for each measure, which coincide with maximum distance separable codes of order k. In addition to deepening the understanding of these measures and multivariate information more generally, we use these results to show that previously proposed bounds on information geometric quantities are tight at the extremes.

show abstract

Prevalence and patterns of higher-order drug interactions in Escherichia coli

Cited by 75 publications

References 40 publications

Characterising a stable five-species microbial community for use in experimental evolution and ecology

Characterising a stable five-species microbial community for use in experimental evolution and ecology

A Consensus Framework Unifies Multi-Drug Synergy Metrics

Maximizing Multivariate Information With Error-Correcting Codes

Contact Info

Product

Resources

About