Reproducibility is a fundamental expectation in science and enables investigators to have confidence in their research findings and the ability to compare data from disparate sources, but evaluating reproducibility can be elusive. For example, generating RNA sequencing (RNA-seq) data includes multiple steps where variance can be introduced. Thus, it is unclear if RNA-seq data from different sources can be validly compared. While most studies on RNA-seq reproducibility focus on eukaryotes, we evaluate bias in bacteria using Pseudomonas aeruginosa gene expression data from five laboratory models of cystic fibrosis. We leverage a large data set that includes samples prepared in three different laboratories and paired data sets where the same sample was sequenced using at least two different sequencing pipelines. We report here that expression data are highly reproducible across laboratories. In addition, while samples sequenced with different sequencing pipelines showed significantly more variance in expression profiles than between labs, gene expression was still highly reproducible between sequencing pipelines. Further investigation of expression differences between two sequencing pipelines revealed that library preparation methods were the largest source of error, though analyses to identify the source of this variance were inconclusive. Consistent with the reproducibility of expression between sequencing pipelines, we found that different pipelines detected over 80% of the same differentially expressed genes with large expression differences between conditions. Thus, bacterial RNA-seq data from different sources can be validly compared, facilitating the ability to advance understanding of bacterial behavior and physiology using the wide array of publicly available RNA-seq data sets.IMPORTANCERNA sequencing (RNA-seq) has revolutionized biology, but many steps in RNA-seq workflows can introduce variance, potentially compromising reproducibility. While reproducibility in RNA-seq has been thoroughly investigated in eukaryotes, less is known about pipelines and workflows that introduce variance and biases in bacterial RNA-seq data. By leveraging Pseudomonas aeruginosa transcriptomes in cystic fibrosis models from different laboratories and sequenced with different sequencing pipelines, we directly assess sources of bacterial RNA-seq variance. RNA-seq data were highly reproducible, with the largest variance due to sequencing pipelines, specifically library preparation. Different sequencing pipelines detected overlapping differentially expressed genes, especially those with large expression differences between conditions. This study confirms that different approaches to preparing and sequencing bacterial RNA libraries capture comparable transcriptional profiles, supporting investigators’ ability to leverage diverse RNA-seq data sets to advance their science.

Our understanding of bacterial physiology during human infection is limited by the difficulty in assessing bacterial function at the infection site. Recent studies have begun to address this question by quantifying bacterial mRNA levels in human-derived samples using transcriptomics.

Laboratory models are critical to basic and translational microbiology research. Models serve multiple purposes, from providing tractable systems to study cell biology to allowing the investigation of inaccessible clinical and environmental ecosystems. Although there is a recognized need for improved model systems, there is a gap in rational approaches to accomplish this goal. We recently developed a framework for assessing the accuracy of microbial models by quantifying how closely each gene is expressed in the natural environment and in various models. The accuracy of the model is defined as the percentage of genes that are similarly expressed in the natural environment and the model. Here, we leverage this framework to develop and validate two generalizable approaches for improving model accuracy, and as proof of concept, we apply these approaches to improve models of Pseudomonas aeruginosa infecting the cystic fibrosis (CF) lung. First, we identify two models, an in vitro synthetic CF sputum medium model (SCFM2) and an epithelial cell model, that accurately recapitulate different gene sets. By combining these models, we developed the epithelial cell-SCFM2 model which improves the accuracy of over 500 genes. Second, to improve the accuracy of specific genes, we mined publicly available transcriptome data, which identified zinc limitation as a cue present in the CF lung and absent in SCFM2. Induction of zinc limitation in SCFM2 resulted in accurate expression of 90% of P. aeruginosa genes. These approaches provide generalizable, quantitative frameworks for microbiological model improvement that can be applied to any system of interest.

Background: Developing therapies for chronic lung infection in people with cystic fibrosis (CF) requires understanding how infectious microbes such as Pseudomonas aeruginosa behave in the CF lung.P. aeruginosa is usually studied in laboratory models, but how well these models mimic the CF lung is not always known.To evaluate P. aeruginosa behavior in CF lab models, a quantitative framework was developed using gene expression to compare models with a benchmark of human infection, leading to the finding that synthetic CF sputum (SCFM2) captures more than 85% of P. aeruginosa genes expressed in the CF lung [1].Here, we ask how the P. aeruginosa strain PAO1 performs in a murine model if grown in SCFM2 before inoculation.Methods: We grew PAO1 in SCFM2 for 6 hours and then extracted ribonucleic acid (RNA) (in vitro samples) or inoculated mice.Female C57BL/ 6 mice were infected intranasally with PAO1/SCFM2 (10 7 colony-forming units (CFU)/mouse).We harvested lungs 18 hours after infection for RNA or CFU counts.Libraries were prepared from total RNA of in vitro samples or lungs with the NEBNext smRNA Library Prep Set before sequencing on Illumina NovaSeq or NextSeq platforms.Reads from our samples and published in vitro samples [2] were mapped to the PAO1 genome, and normalized read counts per gene were used to measure how well PAO1 gene expression from in vitro and lung samples captured P. aeruginosa gene expression in sputum by calculating normalized expression per gene relative to its expression level in published sputum RNA sequencing (seq) data [1].Model accuracy is the percentage of genes with normalized expression within 2 standard deviations of their expression level in sputum.Results: CFU output from infected mouse lungs ranged from 10 7 to 10 8 , and RNA-seq of lungs (200M reads) yielded 260,676 to more than 4 million reads mapping to PAO1 coding sequences.This coverage was sufficient to calculate an accuracy score for our model.In vitro and lung samples were highly accurate, capturing 88.3% and 84.6% of P. aeruginosa genes expressed in sputum, respectively.Of genes not mimicking P. aeruginosa gene expression in sputum, we found some (323) in in vitro and lung data sets, whereas some were unique to in vitro (320) or lung (520) data sets.Conclusions: Deeply sequenced mouse lungs infected with PAO1 have enough reads to analyze PAO1 gene expression, and mice infected with PAO1 grown in SCFM2 reflect P. aeruginosa gene expression in sputum at an accuracy comparable with that of published in vitro samples [1].Numerous genes not mimicking P. aeruginosa gene expression in sputum are unique to in vitro or lung data sets, suggesting that PAO1 in our mouse model may perform better in some functional categories than SCFM2, and vice versa.Furthermore, the PAO1 transcriptome was 4% more accurate in our mouse model than in a previously published murine pneumonia model where PAO1 was grown on a standard growth medium [1,3], raising the possibility that using PAO1 grown in SCFM2 as the inoculum increases accuracy of murine CF models.Work is underway to test this hypothesis.

The role of asymptomatic carriers in transmission poses challenges for control of the COVID-19 pandemic. Study of asymptomatic transmission and implications for surveillance and disease burden are ongoing, but there has been little study of the implications of asymptomatic transmission on dynamics of disease. We use a mathematical framework to evaluate expected effects of asymptomatic transmission on the basic reproduction number R0 (i.e., the expected number of secondary cases generated by an average primary case in a fully susceptible population) and the fraction of new secondary cases attributable to asymptomatic individuals. If the generation-interval distribution of asymptomatic transmission differs from that of symptomatic transmission, then estimates of the basic reproduction number which do not explicitly account for asymptomatic cases may be systematically biased. Specifically, if asymptomatic cases have a shorter generation interval than symptomatic cases, R0 will be over-estimated, and if they have a longer generation interval, R0 will be under-estimated. Estimates of the realized proportion of asymptomatic transmission during the exponential phase also depend on asymptomatic generation intervals. Our analysis shows that understanding the temporal course of asymptomatic transmission can be important for assessing the importance of this route of transmission, and for disease dynamics. This provides an additional motivation for investigating both the importance and relative duration of asymptomatic transmission.

Laboratory models are a cornerstone of modern microbiology, but the accuracy of these models has not been systematically evaluated. As a result, researchers often choose models based on intuition or incomplete data. We propose a general quantitative framework to assess model accuracy from RNA sequencing data and use this framework to evaluate models of Pseudomonas aeruginosa cystic fibrosis (CF) lung infection. We found that an in vitro synthetic CF sputum medium model and a CF airway epithelial cell model had the highest genome-wide accuracy but underperformed on distinct functional categories, including porins and polyamine biosynthesis for the synthetic sputum medium and protein synthesis for the epithelial cell model. We identified 211 "elusive" genes that were not mimicked in a reference strain grown in any laboratory model but found that many were captured by using a clinical isolate. These methods provide researchers with an evidence-based foundation to select and improve laboratory models. IMPORTANCE Laboratory models have become a cornerstone of modern microbiology. However, the accuracy of even the most commonly used models has never been evaluated. Here, we propose a quantitative framework based on gene expression data to evaluate model performance and apply it to models of Pseudomonas aeruginosa cystic fibrosis lung infection. We discovered that these models captured different aspects of P. aeruginosa infection physiology, and we identify which functional categories are and are not captured by each model. These methods will provide researchers with a solid basis to choose among laboratory models depending on the scientific question of interest and will help improve existing experimental models.

How unicellular organisms optimize the production of compounds is a fundamental biological question. While it is typically thought that production is optimized at the individual-cell level, secreted compounds could also allow for optimization at the group level, leading to a division of labor where a subset of cells produces and shares the compound with everyone. Using mathematical modeling, we show that the evolution of such division of labor depends on the cost function of compound production. Specifically, for any trait with saturating benefits, linear costs promote the evolution of uniform production levels across cells. Conversely, production costs that diminish with higher output levels favor the evolution of specialization-especially when compound shareability is high. When experimentally testing these predictions with pyoverdine, a secreted iron-scavenging compound produced by Pseudomonas aeruginosa, we found linear costs and, consistent with our model, detected uniform pyoverdine production levels across cells. We conclude that for shared compounds with saturating benefits, the evolution of division of labor is facilitated by a diminishing cost function. More generally, we note that shifts in the level of selection from individuals to groups do not solely require cooperation, but critically depend on mechanistic factors, including the distribution of compound synthesis costs.

Although bacteria have been studied in infection for over 100 years, the majority of these studies have utilized laboratory and animal models that often have unknown relevance to the human infections they are meant to represent. A primary challenge has been to assess bacterial physiology in the human host. To address this challenge, we performed transcriptomics of S. aureus during human cystic fibrosis (CF) lung infection. Using a machine learning framework, we defined a “human CF lung transcriptome signature” that primarily included genes involved in metabolism and virulence. In addition, we were able to apply our findings to improve an in vitro model of CF infection. Understanding bacterial gene expression within human infection is a critical step toward the development of improved laboratory models and new therapeutics.

Laboratory experiments have uncovered many basic aspects of bacterial physiology and behavior. After the past century of mostly in vitro experiments, we now have detailed knowledge of bacterial behavior in standard laboratory conditions, but only a superficial understanding of bacterial functions and behaviors during human infection. It is well-known that the growth and behavior of bacteria are largely dictated by their environment, but how bacterial physiology differs in laboratory models compared with human infections is not known. To address this question, we compared the transcriptome of Pseudomonas aeruginosa during human infection to that of P. aeruginosa in a variety of laboratory conditions. Several pathways, including the bacterium's primary quorum sensing system, had significantly lower expression in human infections than in many laboratory conditions. On the other hand, multiple genes known to confer antibiotic resistance had substantially higher expression in human infection than in laboratory conditions, potentially explaining why antibiotic resistance assays in the clinical laboratory frequently underestimate resistance in patients. Using a standard machine learning technique known as support vector machines, we identified a set of genes whose expression reliably distinguished in vitro conditions from human infections. Finally, we used these support vector machines with binary classification to force P. aeruginosa mouse infection transcriptomes to be classified as human or in vitro. Determining what differentiates our current models from clinical infections is important to better understand bacterial infections and will be necessary to create model systems that more accurately capture the biology of infection.

Microbes produce many molecules that are important for their growth and development, and the consumption of these secretions by nonproducers has recently become an important paradigm in microbial social evolution. Though the production of these public goods molecules has been studied intensely, little is known of how the benefits accrued and costs incurred depend on the quantity of public good molecules produced. We focus here on the relationship between the shape of the benefit curve and cellular density with a model assuming three types of benefit functions: diminishing, accelerating, and sigmoidal (accelerating then diminishing). We classify the latter two as being synergistic and argue that sigmoidal curves are common in microbial systems. Synergistic benefit curves interact with group sizes to give very different expected evolutionary dynamics. In particular, we show that whether or not and to what extent microbes evolve to produce public goods depends strongly on group size. We show that synergy can create an “evolutionary trap” which can stymie the establishment and maintenance of cooperation. By allowing density dependent regulation of production (quorum sensing), we show how this trap may be avoided. We discuss the implications of our results for experimental design.

Once thought to live independently, bacteria are now known to be highly social organisms. Their behaviors ranges from cooperatively forming complex multispecies communities to fiercely competing for resources. Work over the past fifty years has shown that bacteria communicate through diverse mechanisms, such as exchanging diffusible molecules, exporting molecules in membrane vesicles, and interacting through direct cell-cell contact. These methods allow bacteria to sense and respond to other cells around them and coordinate group behavior. In this review, we share the discoveries and lessons learned in the field of bacterial communication with the aim of providing insights to parasitologists and other researchers working on related questions.

Bacteria often face fluctuating environments, and in response many species have evolved complex decision- making mechanisms to match their behaviour to the prevailing conditions. Some environmental cues provide direct and reliable information (such as nutrient concentrations) and can be responded to individually. Other environmental parameters are harder to infer and require a collective mechanism of sensing. In addition, some environmental challenges are best faced by a group of cells rather than an individual. In this review, we discuss how bacteria sense and overcome environmental challenges as a group using collective mechanisms of sensing, known as 'quorum sensing' (QS). QS is characterized by the release and detection of small molecules, potentially allowing individuals to infer environmental parameters such as density and mass transfer. While a great deal of the molecular mechanisms of QS have been described, there is still controversy over its functional role. We discuss what QS senses and how, what it controls and why, and how social dilemmas shape its evolution. Finally, there is a growing focus on the use of QS inhibitors as antibacterial chemotherapy. We discuss the claim that such a strategy could overcome the evolution of resistance. By linking existing theoretical approaches to data, we hope this review will spur greater collaboration between experimental and theoretical researchers.

Microbes produce many molecules that are important for their growth and development, and the consumption of these secretions by nonproducers has recently become an important paradigm in microbial social evolution. Though the production of these public goods molecules has been studied intensely, little is known of how the benefits accrued and costs incurred depend on the quantity of public good molecules produced. We focus here on the relationship between the shape of the benefit curve and cellular density with a model assuming three types of benefit functions: diminishing, accelerating, and sigmoidal (accelerating then diminishing). We classify the latter two as being synergistic and argue that sigmoidal curves are common in microbial systems. Synergistic benefit curves interact with group sizes to give very different expected evolutionary dynamics. In particular, we show that whether or not and to what extent microbes evolve to produce public goods depends strongly on group size. We show that synergy can create an “evolutionary trap” which can stymie the establishment and maintenance of cooperation. By allowing density dependent regulation of production (quorum sensing), we show how this trap may be avoided. We discuss the implications of our results for experimental design. 4

The Independent Action Hypothesis (IAH) states that pathogenic individuals (cells, spores, virus particles etc.) behave independently of each other, so that each has an independent probability of causing systemic infection or death. The IAH is not just of basic scientific interest; it forms the basis of our current estimates of infectious disease risk in humans. Despite the important role of the IAH in managing disease interventions for food and water-borne pathogens, experimental support for the IAH in bacterial pathogens is indirect at best. Moreover since the IAH was first proposed, cooperative behaviors have been discovered in a wide range of microorganisms, including many pathogens. A fundamental principle of cooperation is that the fitness of individuals is affected by the presence and behaviors of others, which is contrary to the assumption of independent action. In this paper, we test the IAH in Bacillus thuringiensis (B.t), a widely occurring insect pathogen that releases toxins that benefit others in the inoculum, infecting the diamondback moth, Plutella xylostella. By experimentally separating B.t. spores from their toxins, we demonstrate that the IAH fails because there is an interaction between toxin and spore effects on mortality, where the toxin effect is synergistic and cannot be accommodated by independence assumptions. Finally, we show that applying recommended IAH dose-response models to high dose data leads to systematic overestimation of mortality risks at low doses, due to the presence of synergistic pathogen interactions. Our results show that cooperative secretions can easily invalidate the IAH, and that such mechanistic details should be incorporated into pathogen risk analysis.

“He will win who knows when to fight and when not to fight.”― Sun Tzu, The Art of War, circa 500 BC (1) Microbes are remarkably social. They live in complex, interdependent communities where they share and exchange a variety of beneficial compounds ranging from cell−cell signals to iron-scavenging siderophores to digestive enzymes. Cooperation, particularly between cells of a single genotype (2), is fundamental for how cells survive and grow. However, life is not always so amicable. Microbes are regularly confronted by other genotypes armed to the teeth with weapons including secreted toxins, domesticated viruses, and even poisoned spears. Some microbial toxins have been exploited for decades to produce clinical antibiotics. However, several papers have argued that many antibiotics at their low, ecologically relevant concentrations are, in fact, friendly signals that coordinate community functions (3). A study in PNAS by Abrudan et al. (4) challenges this view, focusing on 13 natural bacterial strains from the most famous genus of antibiotic producers, Streptomyces . The authors seek to firmly reestablish antibiotic production and regulation in terms of the logic of attack and defense. Streptomyces strains are well known to inhibit and kill one another, but what determines the strength and impact of these attacks? To address this question, Abrudan et al. (4) studied how antibiotic inhibition from one strain changes when a second strain is nearby. All strains were studied under two conditions: “asocial” and “social” (Fig. 1 B ). The asocial experiments evaluated the baseline tendency of one strain to harm others. This was done by simply growing a focal strain alone on agar for a few days before adding a target strain on top (in soft agar) and asking whether this second strain could grow. The social experiments were different in that the focal strain was grown next to … [↵][1]1To whom correspondence may be addressed. Email: dcornforth{at}gmail.com or kevin.foster{at}zoo.ox.ac.uk. [1]: #xref-corresp-1-1

Quorum sensing (QS) is a cell-cell communication system that controls gene expression in many bacterial species, mediated by diffusible signal molecules. Although the intracellular regulatory mechanisms of QS are often well-understood, the functional roles of QS remain controversial. In particular, the use of multiple signals by many bacterial species poses a serious challenge to current functional theories. Here, we address this challenge by showing that bacteria can use multiple QS signals to infer both their social (density) and physical (mass-transfer) environment. Analytical and evolutionary simulation models show that the detection of, and response to, complex social/physical contrasts requires multiple signals with distinct half-lives and combinatorial (nonadditive) responses to signal concentrations. We test these predictions using the opportunistic pathogen Pseudomonas aeruginosa and demonstrate significant differences in signal decay between its two primary signal molecules, as well as diverse combinatorial responses to dual-signal inputs. QS is associated with the control of secreted factors, and we show that secretome genes are preferentially controlled by synergistic "AND-gate" responses to multiple signal inputs, ensuring the effective expression of secreted factors in high-density and low mass-transfer environments. Our results support a new functional hypothesis for the use of multiple signals and, more generally, show that bacteria are capable of combinatorial communication.

The field of ecology has long recognized two types of competition: exploitative competition, which occurs indirectly through resource consumption, and interference competition, whereby one individual directly harms another. Here, we argue that these two forms of competition have played a dominant role in the evolution of bacterial regulatory networks. In particular, we argue that several of the major bacterial stress responses detect ecological competition by sensing nutrient limitation (exploitative competition) or direct cell damage (interference competition). We call this competition sensing: a physiological response that detects harm caused by other cells and that evolved, at least in part, for that purpose. A key prediction of our hypothesis is that bacteria will counter-attack when they sense ecological competition but not when they sense abiotic stress. In support of this hypothesis, we show that bacteriocins and antibiotics are frequently upregulated by stress responses to nutrient limitation and cell damage but very rarely upregulated by stress responses to heat or osmotic stress, which typically are not competition related. We argue that stress responses, in combination with the various mechanisms that sense secretions, enable bacteria to infer the presence of ecological competition and navigate the 'microbe-kill-microbe' world in which they live.