Time to start taking time seriously: how to investigate unexpected biological rhythms within infectious disease research

Abstract

The discovery of rhythmicity in host and pathogen activities dates back to the Hippocratic era, but the causes and consequences of these biological rhythms have remained poorly understood. Rhythms in infection phenotypes or traits are observed across taxonomically diverse hosts and pathogens, suggesting general evolutionary principles. Understanding these principles may enable rhythms to be leveraged in manners that improve drug and vaccine efficacy or disrupt pathogen timekeeping to reduce virulence and transmission. Explaining and exploiting rhythms in infections require an integrative and multidisciplinary approach, which is a hallmark of research within chronobiology. Many researchers are welcomed into chronobiology from other fields after observing an unexpected rhythm or time-of-day effect in their data. Such findings can launch a rich new research topic, but engaging with the concepts, approaches and dogma in a new discipline can be daunting. Fortunately, chronobiology has well-developed frameworks for interrogating rhythms that can be readily applied in novel contexts. Here, we provide a ‘how to’ guide for exploring unexpected daily rhythms in infectious disease research. We outline how to establish: whether the rhythm is circadian, to what extent the host and pathogen are responsible, the relevance for host–pathogen interactions, and how to explore therapeutic potential.

This article is part of the Theo Murphy meeting issue ‘Circadian rhythms in infection and immunity’.

1. It is about time to uncover the importance of biological rhythms in infectious disease

Biological time is a frequently neglected variable in research, but there is growing recognition of its importance in interactions between hosts, vectors and agents of infectious disease. Historic efforts to understand rhythms in infectious disease were often labour-intensive case studies on individual patients (figure 1a). In the modern era, rhythmic patterns are emerging from advances in ‘omics technologies, real-time imaging platforms and data science (e.g. [2–5]). Their decreasing cost makes round-the-clock examination of infection dynamics more accessible across all levels of biological scale, from molecular to population studies, from the lab to the field. Consequently, reports of temporal phenotypes during infections are increasingly appearing in the literature (figure 1b,c). We note that research on ‘seasonal’ infection and immunity far exceeds that on other kinds of ‘rhythms’. However, seasonal studies tend not to harness chronobiology concepts, which is a missed opportunity to apply this robust framework for exploring oscillatory behaviour that we introduce below.

Figure 1.

Historical and recent medical and scientific relevance. (a) Daily temperature rhythm data from a malaria-infected patient in 1901, collected as part of symptom monitoring to assess antimalarial treatment effects [1] . Upsurge in publications addressing (b) circadian/daily and (c) seasonal rhythms in infection and immunity, highlighting that seasonal rhythms are more often studied. Graphs depict the numbers of studies published each year from 1940 to 2020 (Source: PubMed search, https://pubmed.ncbi.nlm.nih.gov).

Parasites, viruses, bacteria and other microbes (henceforward referred to as pathogens) encounter host environments that change with the time-of- day or season. For example, circadian rhythms have been characterized across the immune systems of diverse taxa (e.g. plants, mammals, insects [6–12]). In parallel, rhythms have been identified in pathogen traits that underpin their virulence and transmission, including metabolism, replication and dissemination throughout tissues [13–15]. Understanding how rhythms shape interactions between hosts, pathogens and vectors, and why pathogens express their own rhythms, is challenging because it requires disentangling oscillations across multiple interacting species in the context of complex pathogen lifecycles (figure 2).

Figure 2.

Schematic of methodologies used to interrogate rhythms in infections. (a) Environmental factors and host–pathogen rhythms explored in this issue: environmental factors influence the intrinsic circadian clocks of hosts or/and pathogens and response systems to the pathogen. These factors regulate rhythmic activities in both host and pathogen (e.g. immune responses, pathogen replication), affecting infection dynamics. Understanding these interactions is critical in the field of chronopathology, which investigates how rhythms in hosts and pathogens impact disease outcomes and treatment effectiveness. (b) Defining rhythmicity: a biological rhythm is characterized by period, phase and amplitude (see table A1 for definitions). Alterations in pathogen rhythms relative to host rhythms can influence infection outcomes over a 24-h cycle. (c) Manipulating rhythms: altering host or pathogen rhythms through methods such as overexpression, gene silencing/knockout (KO) or rhythm inversion can impact pathogen activities (illustrated by mRNA level and viral load; natural rhythm and its impacts in grey, perturbations and consequences in red). (d) Circadian (mis)alignment: differences in circadian alignment between pathogen and recipient hosts can reveal characteristics of pathogen rhythms and their consequences.

The original research and perspectives in this theme issue build upon this rapidly developing field of ‘chronopathology' (chronos = time, pathology = study of disease) and highlight future areas of enquiry (figure 2a). Mechanisms governing daily regulation of host immunity, metabolism and homeostasis are beginning to be elucidated (e.g. Costantini et al. [16]; Hirako et al. [17]; Major-Styles et al. [18]; Mihut et al. [19]; Munns et al. [20]; Rageema et al. [21]), with circadian clocks often determining the magnitude of responses (Paajanen et al. [22]), but why these rhythms exist and their relevance during infection is poorly understood (e.g. Holland et al. [23]; Larrondo [24]). Beyond daily timekeeping, hosts and vectors can exhibit seasonal changes in their activity and responses that could affect disease transmission and population dynamics of infections (e.g. Rund et al. [25]; Sayantan et al. [26]). For pathogens, establishing what drives rhythmic traits, how these rhythms are controlled and how they interact with host timekeeping to impact disease severity and pathogen fitness are major questions (e.g. Greischar and Childs [27]; Hirako et al. [17]; Holland et al. [23]; Larrondo [24]). Given that the time-of-day of infection can dictate life-or-death outcomes for hosts and pathogens [28,29], answers to these questions are urgently needed. Moreover, understanding the roles of rhythms in infections can inform the development new interventions and the time-of-day to deploy them most effectively.

Chronobiology research has generally focused on elucidating timekeeping mechanisms in model experimental species, most often the house mouse Mus musculus, vinegar fly Drosophila melanogaster, flowering plant Arabidopsis thaliana, fungus Neurospora crassa and cyanobacterium Synechococcus elongatus, with the molecular fundamentals of endogenous clocks still debated (e.g. Larrondo [24]; Mihut et al. [19]). However, chronobiology is also being applied to tackle local and global threats to human health and the environment (e.g. Rees et al. [30]; Paajanen et al. [22]). Articles in this issue, for example, address how sleep and its timing corelate with infection susceptibility in humans (Martinez-Albert et al. [31]), the timing of flight in the malaria mosquito vector Anopheles stephensi (Rund et al. [25]) and rhythmic activity of Prochlorococcus marinus, a marine cyanobacterium that constitutes the largest carbon sink on Earth (Peng et al. [32]), as well as explaining the rhythms of pathogens (Hirako et al. [17]; Holland et al. [23]). Expanding our understanding beyond model organisms will not only aid these efforts, but also uncover fundamental aspects of circadian clock evolution and help identify novel rhythms in microbes of clinical and environmental importance (Larrondo [24]).

While chronobiologists set out to test for the presence of rhythms in hosts and pathogens, or in their interactions (e.g. [33–36]), other studies make these observations by a more serendipitous route. For example, the observation that hepatitis C virus (HCV)-infected patients undergoing liver transplantation experienced faster viral replication when transplant surgery occurred in the morning compared with in the afternoon launched the study of how host rhythms regulate flaviviruses [37]. In addition to discoveries stemming from the clinic, time-of-day effects emerge when researchers alter the schedule of frequently conducted experiments. For example, they may notice that a disease phenotype is more severe when hosts are infected at a certain time- of-day, or that unexpected results stem from sampling at a different time-of-day (e.g. [38,39]). Alternatively, so-called ‘clock genes’ may stand out as important interacting partners for pathogen proteins during infection (e.g. [40,41]).

Here, we guide researchers taking forward hypotheses or observations of temporal phenotypes during infections. Our focus is circadian clocks but the tools and concepts extend to biological rhythms with seasonal, infradian (>24 h) and ultradian (<24 h) durations. Engaging with the core concepts and approaches in a new discipline can be daunting, especially when changing gears from conducting lab work at a single biological time- of-day to round-the-clock studies. The substantially increased cost—both to grant budgets and to the sanity of researchers—requires robust identification of key questions and rigorous experimental design. We wish to equip researchers with the knowledge, conceptual framework and confidence to take their research into a new paradigm while avoiding common chronobiology pitfalls. Our ‘hitchhiker’s’ guide aims to: (i) explain why rhythms in infections matter, and what may be gained from considering time-of-day; introduce core chronobiology concepts and how these translate to infectious diseases; (iii) outline different approaches to study rhythms in infections and their application to diverse study systems; and (iv) explore how explaining rhythms in infections can help solve real-world problems. Throughout, we illustrate experimental (figure 2b–d) and analytical (figure 3) aspects of working on rhythms, and our glossary (table 1 in appendix A) provides a practical guide that in addition to defining terms, highlights common misconceptions and how to avoid them. While we focus on infectious diseases, our template applies to any study system or topic meeting chronobiology for the first time.

Figure 3.

Visualization of rhythms. Appropriate choice of figures depends on the type of data, how it has been processed and the biological insight being illustrated. Indicating when data are collected from day (light) and night (dark) is usually denoted by white-black bars (a,b,d), or with grey shaded regions to indicate dark (d–f). It is good practice to include raw data (a) and this format aids comparison of the phase and amplitude of different groups, but for visualizing many groups, a heatmap is useful (b) although while rhythmicity per se is visible, resolution of phase and amplitude may be reduced. The intensity and amplitude of measurements from cell-reporters can decay over time, requiring detrending and normalizing to deconvolve the biological rhythm of interest from other shorter- and/or longer-term trends (c). When data are from higher resolution time series than in (a), the raw traces are often illustrated to give an indication of noise in the dataset (d). Behaviour data are commonly plotted as an actogram (e), where each row is a different day (i.e. circadian cycle) and double plotting allows each cycle to be compared with the next, with the height of bars indicating activity level. Actograms are particularly useful for illustrating the effects of perturbating an environmental rhythm, which is indicated by moving the white and grey shading between rows (here, a phase shift of 6 h which advances the rhythm is illustrated). Data illustrated in (c–e) represent continuous or repeated measures on an individual independent replicate (e.g. a single organism or culture) and can confer higher temporal resolution, less variance and more power to analyses than data represented in (a,b), which are often derived from destructive sampling so that each time point represents different replicates. Polar or Rayleigh plots (f) are particularly good for visualizing phase because the cyclical nature is clear. In addition to displaying the rhythm and its characteristics, it is good practice to give an indication of the strength of support for the rhythm, so periodograms (g) are often included in electronic supplementary figures.

2. Conceptual toolkit

Circadian clocks coordinate organisms’ activities with daily environmental rhythms, enabling them to cope with the predictable consequences of the Earth’s solar rotation. Thus, life in a rhythmic world dictates the timing of activities for many organisms. This includes pathogens, for which the environments within their hosts and vectors change dramatically over 24 h, in fundamental cellular processes (e.g. transcription, translation, bioenergetics), systemic physiology and homeostasis (e.g. metabolism, body temperature, immune responses) and whole-organism behaviours (e.g. locomotor activity, reproduction, foraging). These myriad daily rhythms of hosts and vectors offer opportunities for pathogens to exploit and they present dangers to evade at certain times of day [13,15,29,42]. Vice versa, hosts and vectors can deploy their rhythms, or may need to overrule certain rhythms, to evade and defend against pathogens [43]. Given that hosts (including vectors) become infected at specific times of day—often driven by rhythms in foraging—and that rhythms govern infection progression, researchers may wish to elucidate:

1. What genetic and molecular mechanisms allow host circadian clocks to influence responses to infection? Is it beneficial to fine-tune this control in response to different pathogens?

2. Do pathogens possess circadian oscillators of their own and what drives them? Or are they intrinsically arrhythmic and allow host rhythms to enforce rhythmicity—for example, by killing mis-timed pathogens or reliance on the host’s rhythmic cellular machinery for replication and dissemination?

3. What are the opportunities and dangers that host rhythms present to parasites and how do pathogens exploit/evade them, respectively?

4. Can pathogens determine the time-of-day when they infect and manipulate host circadian rhythms to their own ends? Do host and pathogen rhythms coevolve?

5. Is there feedback in which circadian rhythms and immune responses both influence each other? Do they interact to balance resistance against pathogens and tolerance to the costs of infection, and to aid recovery?

Answering these questions requires a framework grounded in well-verified chronobiology concepts. For example, a phenotype that depends on time- of-day does not necessarily constitute a circadian rhythm or reveal an endogenous oscillator, but this is a plausible place to start given that circadian timekeeping permeates biology and was likely an emergent facet of cells (table 1 in appendix A).

(a). What are circadian rhythms?

Circadian rhythms are approximately 24 h oscillations in behaviour, physiology and cellular activity that are exhibited by most organisms. They rely on intrinsic timing mechanisms and are distinguished from passive responses to environmental fluctuations by three primary characteristics [44,45]. First, circadian rhythms are self-sustaining and persist in constant conditions when no external timing cues are received. Second, circadian rhythms can synchronize their phase to external timing cues (‘zeitgebers’), such as daily light/dark and feed/fast cycles. This entrainment locks the intrinsic oscillation to exactly 24 h and ensures that the rhythmic phenomenon occurs at the appropriate time-of-day, often allowing biological oscillations to functionally anticipate 24 h environmental cycles. Finally, the speed of the rhythm—the period—does not vary over biologically relevant temperature ranges, ensuring that timekeeping is compensated against temperature fluctuations. By contrast, daily rhythms can be generated by a direct response to cyclical environmental stimuli, but these are not anticipatory or endogenous so the same response is mounted independently of the time-of day at which the stimuli occurs. Organization of circadian rhythms, their oscillatory parameters (figure 2b) and how they temporally regulate, or ‘gate’, cellular physiology are discussed in detail in this theme issue (e.g. Mihut et al. [19]; Paajanen et al. [22]).

Daily oscillations are observed at all levels of biological scale, but circadian rhythms are a cellular phenomenon: cells have an endogenous molecular clock that allows them to keep track of time. The canonical model involves delayed transcription–translation feedback loops (TTFL), whereby transcriptional activators drive expression of so called ‘clock genes’ that encode products that directly or indirectly repress their own transcription to yield a 24 h oscillation, which facilitates rhythmic expression of ‘clock-controlled genes’ (CCG). The clock gene circuitry has been characterized in every eukaryotic lineage but is poorly conserved. Thus, absence of canonical TTFL or clock genes in a pathogen that displays a rhythmic phenotype cannot be interpreted as evidence for the absence of a circadian oscillator.

Mammalian circadian clock protein complexes of the TTFL function in concert with global histone modification and chromatin remodelling to generate rhythmic gene transcription, with ca 40% of protein-encoding transcripts oscillating in at least one tissue in mice and humans. However, there is a disparity between rhythmic transcripts and rhythms in protein abundance, with often little correlation between the two [5,46–48]. This challenges the linear narrative that TTFL-driven mRNA oscillations lead to rhythms in protein abundance and activity that underpin circadian regulation of cell physiology. Most protein activity is governed post-translationally, and there is growing recognition of the important role that highly conserved kinases and phosphatases play in driving circadian rhythms (e.g. CK1d/e, GSK3a/b and PP1) [19]. In addition, circadian rhythms in post-translational modifications, ion flux, metabolism and protein synthesis occur in many models where the TTFL is non-functional [3,49–53]; O-], suggesting that entirely post-translational molecular timekeeping is feasible in eukaryotes, as demonstrated for cyanobacteria and the FRQ-less oscillator in Neurospora [54,55]. Although a more complex picture of the circadian clock is emerging, the output remains the same: most fundamental cellular processes robustly oscillate over the course of 24 h, under laboratory and real-world conditions.

The take-home message is that time-of-day differences in host or pathogen phenotypes may be driven by endogenous biological timekeeping in one, several, or none of these parties. In addition, circadian rhythms at the cellular, systemic or behavioural level of the host may all contribute to driving a time-of-day dependent pathogen phenotype. Discriminating between all these possibilities is not a trivial undertaking, but knowing which rhythms matter and how they are generated is key to whether they can be targeted to improve host health or reduce pathogen fitness.

(b). Why possess a circadian clock?

Aligning activities appropriately to daily environmental cycles of light/dark, temperature, humidity and UV exposure is an intuitive way to maximize survival and reproduction (i.e. Darwinian fitness) [56]. For example, undertaking risky foraging behaviour when most predators are resting provides a form of ‘extrinsic adaptive value’. Such extrinsic benefits arise from alignment to environmental rhythms, which may be abiotic or belong to other organisms. Organisms may also garner intrinsic benefits from temporally compartmentalizing incompatible internal processes [57]. A classical example of this is DNA replication being undertaken at the time-of-day that minimizes damage from metabolism-induced oxidative stress or UV exposure. In the context of intrinsic benefits, the time-of-day that activities are undertaken does not matter per se; having a clock whose time is set by an external rhythm provides a convenient and accurate way to organize internal processes. Furthermore, owing to their self-sustaining nature, circadian clocks confer the ability to anticipate and therefore, prepare for, environmental change. For instance, plants modulate the degradation rate of starch during the night so it is optimally depleted according to when they anticipate dawn and the onset of photosynthesis to replenish reserves [53,58].

Despite the ubiquity of circadian clocks across diverse taxa and their assumed fitness benefits, chronobiology has focused on uncovering the molecular mechanisms driving rhythms rather than their ecological and evolutionary explanations. However, the fitness benefits of biological rhythms have been demonstrated for cyanobacteria, malaria parasites, mice, insects and Arabidopsis [34,59–63]. The most commonly used experimental paradigm is to compare proxies for fitness (e.g. replication rate, survival, growth, competitive ability or reproduction) between individuals raised in an environmental cycle that matches their endogenous clock, revealing a competitive advantage over individuals whose clock does not resonate with the duration (period) or timing (phase) of the environmental rhythm (figure 2b). For example, in a semi-wild release study, mutant mice with a shorter (<24 h) circadian period length had lower survival and fecundity compared with those with near 24 h periods, and consequently the prevalence of the ‘short-clock’ gene mutation in the population dropped from ca 50% to 20% within 14 months [61]. Further support for the importance of resonating with environmental rhythms comes from cyanobacteria: long-term evolution in a non-rhythmic environment (constant light) favours round-the-clock photosynthesis and metabolism, which is achieved by mutations that break circadian clockwork [64]. However, while intuitive ideas about the biological function (i.e. fitness costs/benefits) of rhythms in various host and pathogen traits are easy to construct, natural selection is complex and often counter-intuitive, and hard to test in rigorous manner [28,29].

3. Choose your own temporal adventure

This theme issue demonstrates the breadth of methodology and experimental systems used to investigate rhythms during health and disease (figure 2a). It would take several volumes to comprehensively cover all chronobiological tools and techniques used in in vitro and in vivo lab experiments, field studies, computational models and evolutionary theory. Instead, we present a series of case studies that proceed from different starting points.

(a). A screen has identified a ‘clock gene’. Is the infection model circadian-regulated?

Many researchers consider timekeeping upon encountering ‘clock genes’, designated owing to their involvement in the transcriptional–translational feedback loops that regulate some rhythmic gene expression. Perhaps a screen yielded direct interaction between a pathogen and a clock gene product, or revealed differences in their abundance, post-translational modification or rhythmic parameters (amplitude, period, phase; figure 2b). Clock gene products are transcriptional activators, for example CLOCK and BMAL1 in mammals, or repressors such as the PERIOD and CRYPTOCHROME proteins. These proteins are often promiscuous binders with additional functions beyond circadian or transcriptional regulation [18–20] , so may affect the pathogen or host response in a ‘non-circadian’ capacity. Alternatively, the pathogen may specifically utilize a clock protein during replication to harness or manipulate host rhythms to their own end, including replication, as proposed for CLOCK/BMAL1 and herpesvirus interactions [33,40,65]. Elucidating the molecular interactions between a pathogen and circadian clockwork can be informed by asking the following questions.

(i) Are there measurable differences in infection processes depending on the time-of-day of infection?

It is essential to examine this question over several biological scales. For instance, pathogen replication may vary, with specific times of day promoting more aggressive or restrained replication. Furthermore, the time-of-day of infection could influence both local and systemic host responses, including the production of cytokines, the activation of immune cells and the broader inflammatory response. These differences can have knock-on consequences for disease severity, host vulnerability and recovery. Finally, the time-of-day of infection can affect pathogen transmission potential by influencing factors such as dissemination, host behaviour and interactions with other hosts or vectors (e.g. [33,66,67]).

To differentiate between diel and circadian drivers, experiments can be undertaken in constant conditions to infect hosts at different circadian times of day (table 1 in appendix A), while also controlling for any pathogen rhythms (figure 2).

(ii) Does genetic or pharmacological disruption of clock protein function impact upon infection?

Exploring the impact of genetic or pharmacological disruption of circadian clock components involves several considerations. Comparing the effects of conditional versus constitutive knockouts of circadian clock genes can reveal how these genetic alterations influence the immune response, pathogen replication and disease severity. The effects of acute reduction of clock protein function—achieved through methods such as siRNA or CRISPR/Cas9-mediated knockdown—should be contrasted with the consequences of stabilizing these proteins, for example, through pharmacological inhibitors or overexpression. Furthermore, investigating how post-translational modifications of clock proteins, such as phosphorylation or ubiquitination, affect their activity and consequently the host response to infection can provide insights into the dynamic temporal regulation of these proteins. Alterations in these modifications, potentially induced by pharmacological agents, could significantly alter the course of infection. Ultimately, if a clock gene product is critical to an infection process, testing this with a targeted therapeutic may seem enticing. The challenges and limitations of this approach, including how to probe for a bona fide circadian impact by avoiding off-target effects and identifying specific mechanisms of action, are discussed in detail in this issue [20].

(iii) Does infection alter host rhythms?

Understanding whether infection alters host circadian rhythms requires analysis at multiple scales, covering the molecular, physiological and behavioural levels [68]. At the molecular level, infection may disrupt the expression of core clock genes and proteins, which can be monitored using techniques such as bioluminescence assays, or by transcriptomic and proteomic studies of samples collected from circadian-entrained cells or animals [41,69,70]. At the physiological level, infection might influence circadian-regulated processes such as hormone secretion (e.g. cortisol, melatonin), body temperature or metabolic activities, either directly or as a secondary consequence of the immune response [70–72]. Infection may alter circadian-regulated behaviours, such as sleep–wake cycles, feeding patterns or general activity levels, with changes potentially observable through activity monitoring [73,74]. These disruptions across multiple levels would indicate a significant interaction between infection and the host circadian system.

(b). A rhythmic host process impacts my pathogen. Which host rhythm(s) is important?

Host immune responses to infection at the cellular, tissue and systemic levels are often gated by circadian clocks. Intuition suggests that because immune responses defend against pathogens, clock control of immune responses must cause pathogens to be killed in a time-of-day specific manner. Indeed, anti-infection defences are thought to peak during the host’s active phase because this is when pathogens (especially if food-borne) are most likely to be acquired. However, whether immune rhythms are particularly important for standing defences is poorly understood and other aspects of host rhythms might have more impact than foraging on infection risk. For example, some of the deadliest infections are vector-borne and acquired in the host’s rest phase and respiratory viruses are acquired when hosts aggregate, which can be in either the rest or active phase, depending on when the host species has most social interactions. Mechanisms governing rhythms in immune responses and the consequences of these rhythms are relatively well understood, especially for components of the innate response and non-infectious diseases [6,8,10–12,21,43]. Asking how immune rhythms affect infection processes requires considering to what extent an immune factor is regulated in a rhythmic manner to prevent immunopathology, be energetically efficient or for other intrinsic benefits such as temporally compartmentalizing repair and pathogen attack.

Pathogens also rely on their hosts for resources required to replicate and, in the case of viruses, the machinery to synthesize their proteins and assemble new particles for onwards transmission. Given the proportion of metabolites and proteins with differential abundance and activity over 24 h and the myriad cellular processes gated or licenced by the circadian clock [2,19,22], it is likely that a critical resource, protein or process is more accessible at one time or another over 24 h [14,17]. In some scenarios, for example the rhythmic expression of proteins on the cell surface that enable virus binding and entry, maxima and minima of the rhythm substantially alter infection outcomes [75,76]. In other cases, while a particular process may be limiting at certain times of day (especially during sickness), pathogens can reprogramme cells to promote replication, with hosts responding to limit resources and pathogen dissemination [15,33,76]. In these more nuanced instances, any initial time-of-day difference may amplify or diminish as infection progresses. Simultaneously monitoring host rhythms and infection progression presents a challenge because both are dynamic processes.

Ultimately, to establish whether a host rhythm underpins the time-of-day effect on a pathogen, the important questions to answer are the same as for §3a. Altering expression of the rhythmic protein, availability of the rhythmic metabolite or activity of the rhythmic process only tells you whether its presence required for infection, not whether its rhythmicity is required. To demonstrate this, the rhythm must be ‘flatlined’ or inverted (figure 2c), or host and pathogen rhythms misaligned (figure 2d).

(c). A pathogen exhibits a rhythmic phenotype. Does it have a clock?

Do not assume that time-of-day dependent phenotype is the result of a genetically encoded, endogenous pathogen clock because environmental and host factors can drive pathogen rhythms, discussed previously in §3a. The general TTFL structure is found across the kingdoms of life but the molecular components are not conserved across taxa, whereas post-translational regulatory processes and kinase/phosphatase involvement are conserved, so all is not lost in the absence of defined canonical ‘clock genes’. The challenges of identifying timekeeping mechanisms de novo in non-model organisms—including insights, pitfalls and caveats—are discussed in detail in this issue [24]. Establishing the existence of an independent endogenous circadian clock is especially difficult for non-model pathogens [77,78] and appropriate approaches depend on the following:

1. Can the pathogen survive outside the host in constant environmental conditions for a long enough period to observe rhythmicity (ideally >72 h)?

2. What environmental cue(s) synchronize or entrain the pathogen’s rhythms?

3. What can be measured to gauge rhythmicity and for intracellular pathogens, would this output from a putative clock function outside the cellular environment?

How questions (i) and (ii) are tackled depends on the biology of the focal pathogen, but for ideas see [77–80]. Furthermore, several robust and independent oscillatory outputs may be required (e.g. transcriptional, metabolic, post-translational).

Whether this avenue is worth pursuing will depend on how important it is to establish the existence of an endogenous, self-sustaining and temperature-compensated oscillator in the pathogen: trying to conclusively prove the converse—that there is no clock—is extremely challenging. Furthermore, some oscillators, especially if evolutionarily ancient, may only fulfil a subset of criteria. Instead, ask why you need to determine the mechanism of pathogen rhythmicity and whether it is self-sustaining or an environmental response; it is still encoded by pathogen genes— and targetable by medical interventions.

(d). A pathogen exhibits a rhythmic phenotype. Why has this evolved?

While pathogens are sheltered from rhythms in the abiotic environment when living inside others, they are exposed to biotic rhythms belonging to their host/vector. Rhythms in the biotic environment dictate the time-of-day when nutrients (e.g. when food is acquired) and other resources (e.g. ligands used for invasion, replication machinery) are available to pathogens, when dangers are present (e.g. the migration of immune cells) and activity/rest cycles determine when transmission opportunities occur (e.g. the generally nocturnal biting activity of malaria-mosquito vectors [25]). By knowing whether a biotic environmental rhythm presents a danger or an opportunity to a pathogen, one can predict how the pathogen should respond with its own rhythmic strategy (e.g. [23]).

The phase of a pathogen’s rhythm(s) might confer fitness benefits via maximizing the exploitation of host resources or via evading dangers, and pathogens that replicate within the host may also garner fitness benefits from synchronizing their activities with each other [81]. For example, synchronizing waves of replication may enable closely genetically related pathogens to overwhelm host defences, communicate to co-ordinate their developmental decisions or manipulate host rhythms to their own end. However, explaining why a pathogen exhibits a rhythm requires more than an intuitive link between pathogen fitness and an environmental rhythm. This is very challenging when the mechanism driving the pathogen’s rhythm is unknown because perturbating it may not be possible. Assuming it is possible to perturb the pathogen’s rhythm and measure impacts on pathogen fitness (or proxies), asking why a pathogen exhibits a rhythm requires considering the following questions.

(i) Why does timing matter?

Ideally, one perturbs the phase of the pathogen’s rhythm(s) relative to the host rhythm hypothesized to determine pathogen fitness and observes costs to the pathogen (e.g. [82]). This is best achieved by exposing a pathogen to its Zeitgeber or time cue in a manner decoupled from the fitness-determining host rhythm to ‘trick’ the pathogen into displaying a suit of different rhythms while the important within-host environmental rhythm is held constant.

(ii) Are other aspects of the pathogen’s rhythm important?

That pathogens should undertake activities at the time-of-day that maximizes their fitness is obvious, and so, natural selection is expected to act on the timing (phase, figure 2a) of pathogen rhythms. Yet, for pathogens that replicate within the host, the degree of synchrony (measurable via the amplitude) in the activity of individual pathogen cells is also subject to selection. For example, independently of the phase of the pathogen activity, high synchrony could benefit pathogen fitness if collective action facilitates host exploitation. Alternatively, high synchrony in rhythmic replication could be costly if it causes closely (clonally) genetically related pathogens to inadvertently compete for time-limited resources. In cases where extremely high synchrony is costly, pathogens could evolve to be less synchronous via variation in phase and/or period, resulting in different combinations of phase, period and amplitude being equally fit (figure 2). These combinations might be exhibited across individual pathogen cells within an infection (plasticity, see §3d(iv)), or different pathogen genotypes or species might evolve different combinations. Testing these hypotheses is not straightforward because the expansion and contraction of pathogen numbers during infections can cause intuitive measures of synchrony and replication rate to be extremely biased, as explored in detail in this issue (see [27]).

(iii) Does the pathogen gain intrinsic benefits of rhythmicity?

If a pathogen maintained in an arrhythmic environment requires a rhythmic time signal to maintain its fitness, this suggests the pathogen uses the within-host environmental rhythm to set the time of compatible (or not) activities. Thus, the host’s rhythm provides convenient timing information rather than being a selective driver of the pathogen’s rhythm.

(iv) Does the pathogen’s rhythm alter as infections progress, or in different parts of its lifecycle?

Infections are a set of dynamic processes, which may mean that different within-host environmental rhythms are important at different stages of disease progression (e.g. acute versus chronic stages, in the host or vector) and pathogens actively alter their rhythms to suit. Further complexity arises when pathogens express different rhythms at different points in their lifecycle. For example, while rhythmic replication in the blood of the vertebrate host is thought to maximize the within-host fitness of malaria parasites, it is possible that rhythmic replication is also (or alternatively) phased to align the infectiousness of transmission stages to the nocturnal biting of Anopheline mosquito vectors [14]. This highlights that examining the consequences of a pathogen rhythm in only part of its lifecycle can give a misleading view of its adaptive value and selective drivers, as explored in this issue [23].

Overall, pathogen rhythms may be shaped by selective forces associated with the benefits and costs of aligning with multiple rhythms in their biotic and abiotic environments, the intrinsic benefits of rhythmicity and the potential for phase and synchrony to affect fitness independently. This means a pathogen’s rhythm might not appear optimal with respect to one aspect of its rhythm, or perfectly aligned to a specific host rhythm, because it is a compromise of multiple evolutionary pressures, past and present.

4. Analysis toolkit: dealing with time series data

Measuring traits around the clock produces time series datasets in which a repeating ‘circular’ signal(s) is tested for, rather than a single trend across the whole dataset (figure 3a). Finding such patterns amongst noise can be difficult so a robust workflow requires a good dataset, establishing whether there is a rhythmic signal, describing the parameters of the rhythm and then, if applicable, comparing those parameters between treatment groups, before displaying findings (data visualization; figure 3). For most purposes, an ideal dataset has regular (every 2–4 h) equidistant datapoints, spanning at least twice the length of the predicted cycle (e.g. 48 h for a circadian rhythm), with independent replicates [83,84]. Collecting time series from the same individuals over multiple circadian cycles is also desirable, such as when additional dynamic variables that have associated rhythms can be measured (e.g. [23]) or generating expensive ‘omics data. Equidistant datapoints are the main priority because many statistical algorithms to assess rhythmicity (such as JTK_CYCLE, ARSER, RAIN [85–87]) fail without these. For reasons of financial costs, operational feasibility and tractability of infections, compromises between the number of time points and replicates often must be made; here equidistant samples across two or more cycles are better than more samples within one cycle [84]. However, for detecting rhythms in situations where the number of pathogens or immune cells changes during sampling, for example, uneven sampling intervals can avoid overestimating amplitude [88,89].

Analysing time series data is at a minimum a two-step process of determining whether a significant rhythmic signal exists before then estimating the parameters of the rhythm (period, amplitude, phase; figure 2a). Algorithms to test for rhythmicity are broadly categorized as curve-fitting methods: generating curves that vary in shape and assessing how closely each curve fits the original data (e.g. COSOPT [90], harmonic regression, Fast Fourier Transform Nonlinear Least Squares (FFT-NLLS)), or spectral based methods that extract an entire spectrum of signals in the data and determine which is the strongest (e.g. Lomb–Scargle periodogram, Maximum Entropy Spectral Analysis (MESA)). Other approaches include non-parametric tests (e.g. JTK_CYCLE and RAIN) that are widely used for ‘omics datasets’. Choosing between different methods (with disparate statistical origins) may seem overwhelming but this variety benefits researchers because the strengths and weaknesses of all approaches differ, and it is good practice to use a combination of methods to confirm rhythmicity [91]. For example, spectrum-based methods (e.g. MESA) are excellent at estimating parameters but will do so for any signal in the data, regardless of whether it is biologically or statistically significant. It is important to use these algorithms in conjunction with others that provide a level of confidence that there is an underlying rhythm (harmonic regression, FFT-NLLS, JTK_CYCLE, RAIN).

Analyses can be undertaken using several tools for time series data (e.g. [92], BioDare2, CIRCADA [93], PyBoat, CircaCompare, metacycle, Rethomics ([92]| Actimetrics, n.d.; [94], n.d. [11,91,95–97])). For its user-friendliness, we recommend the web-based BioDare2 [91,98]. BioDare2 performs detrending (which helps in determining rhythmicity in data where a pattern might be masked by linear trends such as pathogen replication; figure 3c), includes six methods for parameter estimation and provides visualization of the dataset and of parameter estimates. BioDare2 is also an open data repository, facilitating open science and collaboration. The methods implemented in BioDare2 are particularly powerful for estimating period length, which is one of the more difficult parameters to estimate accurately [91]. However, while BioDare2 can analyse any data (e.g. luminescence, gene expression, pathogen density), we recommend companion software such as Clocklab [92] (Actimetrics, USA), or the R package Rethomics [95] for behavioural (e.g. host activity, figure 3d) data. Rethomics creates plots traditionally used to visualize activity (e.g. actograms; figure 3e) and calculates measurements specific to behavioural data such as the duration and onset of activity bouts. For datasets with very complex dynamics that obscure rhythms, typical of infections, bespoke statistical models or simulations may be necessary [27].

For experiments whose data are solely used to test for, and characterize, a rhythm, no further analysis is necessary. However, for experiments designed to compare rhythms across treatment groups or categories (e.g. figure 3a) the next step is to compare the proportion of replicates that are rhythmic and the parameters of those rhythms between groups. Standard statistical methods are deployed, such as generalized linear models (GLMs) with mixed effects if using groups that have non-independence due to repeated measures from replicates. Note that phase is a circular statistic (i.e. the end of a circadian cycle is also the start of the next cycle), necessitating approaches such as Bayesian circular GLMs [99]. Alternatively, all-in-one curve fitting and parameter estimating methods such as CircaCompare and DoDR can be used to compare rhythmicity across groups [11,100].

Regardless of what an experiment was designed to test and its outcomes, data visualization is essential. How rhythms and their parameters are best illustrated (figure 3) depends on the type of data (e.g. an actogram (figure 3e) for locomotor activity, or a heatmap (figure 3b) for gene expression) and the hypothesis being tested (e.g. change in a rhythm over time, or comparison between treatment groups). Given the circular nature of time-series data, polar plots appropriately display phase and amplitude in an easy-to-understand format similar to a wall clock (except over 24 h instead of 12; figure 3f). Periodograms (figure 3g) provide valuable information on the strength of a rhythm as well as its period, and often form an essential aspect of supplementary information in publications.

5. Translational potential

Circadian rhythms underpin homeostasis and health, and the detrimental consequences of circadian disruption caused by modern lifestyles are well-recognized. Many people will have experienced the negative effects of ‘jet lag’, where the internal phase of their body clocks mismatches with their new time zone. Furthermore, circadian misalignment under controlled laboratory conditions disrupts metabolic homeostasis and immune function in human subjects after only a few days. Prolonged and repeated circadian desynchrony during shift work is associated with severe morbidities including cancer, obesity and type 2 diabetes, and loss of robust circadian timekeeping is a feature of ageing and neurological disorders [101]. In addition to clinical therapeutics, combating disease vectors, crop security and preventing antimicrobial resistance may all benefit from considering time-of-day, as discussed in detail in this theme issue [30].

Protecting host rhythms during infections has great potential to alleviate symptoms for humans as well as other animals. Reinforcing host rhythms may also promote recovery from infection, analogous to the detrimental impact of rhythm disruption on recovery from trauma [102]. Host rhythms might be harnessed to make therapeutic interventions more effective, for example vaccine efficacy may depend on the time-of-day of administration [103]. That >50% of the best-selling drugs in the USA have targets that cycle in a circadian manner highlights the value of at least taking time-of-day into account when treating infections [104,105]. Administering a drug at a particular time to synergize with rhythmic immune responses could enhance efficacy or allow lower doses to be used.

Similarly, pathogens' rhythms affect their fitness, so explaining how their rhythms are controlled and why they are beneficial should stimulate novel approaches for reducing disease severity and transmission. For example, the ability of malaria parasites to slow their replication rhythm confers tolerance to antimalarial drugs [106,107]. Drugs that disrupt pathogen rhythms or take advantage of rhythmic vulnerabilities in pathogen replication may make treatment more effective, especially for drugs with short half lives. These possibilities highlight the need for better translation of chronobiology within and beyond academia, especially with policymakers and stakeholders, as championed by the BioClocksUK community [30].

6. Conclusions

Observations that pathogens express rhythmic behaviours date back to the Hippocratic era and experiments in the 1960s showed that disease severity is a function of the time-of-day of infection, yet general explanations for these phenomena are lacking. Our understanding of the roles of rhythms during infection is still in its infancy, but this novel area of biological enquiry is tractable and translatable. Over the past few decades, chronobiology has focused on ‘what makes a clock tick’ so the time is right to disseminate this knowledge into other disciplines. The role of biological timekeeping during infection has been neglected because different aspects of the underlying biology are studied by disciplines (chronobiology, immunology, evolutionary ecology, parasitology, microbiology, plant sciences, medicine) that often progress independently. While the complexities of considering the rhythms of all parties involved in infections brings new challenges to chronobiology, this topic offers a rare opportunity to explain phenomena across scales of biological organization. We hope the expansive science presented this theme issue, and our guide to getting started, will inspire readers to incorporate chronobiology into their research using a multidisciplinary approach.

Appendix A

See table 1

Table 1.

Consensus chronobiology terminology with common confusions and contentions.

term	definition	useful notes	misconceptions/clarifications
amplitude	The difference between the peak and the average level of a rhythmic process	amplitude reflects the intensity of a rhythm, or the synchrony of rhythms of cells in a sample	misconception: higher amplitude equals a stronger circadian rhythm clarification: amplitude indicates the magnitude of temporal variation, but a higher amplitude does not necessarily indicate a healthier or stronger rhythm
arrhythmic	no evidence of regularity in the pattern of occurrence.	a pattern might not exhibit 24 h periodicity but might be rhythmic with a much shorter period (e.g. an ultradian rhythm), or a longer duration (e.g. lunar and seasonal rhythms)	misconception: arrhythmicity can be concluded from data analysi. clarification: confidence in categorizing arrhythmicity depends on the parameters in the analysis, including the range of periods that a cycle can take, the number of harmonics allowed and the p-value cut-off
biological clock	an internal timekeeping mechanism that generates and regulates the timing of biological rhythms	biological clocks are the systems or mechanisms (often molecular and neural) that generate rhythms	misconception: biological clock and biological rhythm are the same clarification: a biological clock is the underlying timekeeping mechanism, while a biological rhythm may or may not be the output created by a clock
biological rhythm	any repeating physiological or behavioural cycle, not necessarily endogenous.	the observable patterns that clocks produce (e.g. sleep–wake cycle, body temperature fluctuations) that can exist at multiple timescales, not just the 24-h circadian rhythm.
central and peripheral clocks	in mammals the suprachiasmatic nucleus (SCN) is an important pacemaker, contributing to the timing of other (peripheral) clocks within an organism	a hierarchy of clocks—that was applied to mammalian circadian organization—is more complex than previously assumed.	misconception: the SCN sets the time for all other clocks within an organism clarification: peripheral clocks have more autonomy that previously thought and some bi-directional feedback between the SCN and other clocks occurs
circadian clock	a self-sustaining oscillator that allows an organism to exhibit endogenous 24 h rhythms at the molecular, cellular, physiological, behavioural or population levels	the transcription–translation feedback loop (TTFL) is most commonly studied but recent work is revealing that evolutionarily older oscillator mechanisms exist	misconception: all circadian clocks use the same mechanism clarification: different taxonomic groups use different genes to run their TTFL, and non-transcriptional and post-translational oscillators also exist
circadian time (CT)	notation to indicate the internal clock's timekeeping, referring to time-of-day observed in the absence of external cues.	following a switch from 12 : 12 light : dark (LD) conditions to constant dark, CT0−12 represents subjective day (equivalent to night in LD), while CT12−24 represents subjective night (equivalent to day in LD).	misconception: CT is synchronized to the 24 h day clarification: CT reflects the organism's internal rhythm, which might slightly differ from 24 h
clock gene	component of the molecular clock mechanism important for overt circadian rhythm(s), which often feeds back to affect its own expression.	absence or mutation leads to either arrhythmicity or altered overall circadian function; clock genes may also have additional functions beyond circadian regulation	misconception: all genes showing circadian patterns are clock genes clarification: clock genes are those that directly regulate the circadian rhythm, not just any gene with circadian expression patterns
clock mutant	an organism with a mutation in one or more clock genes, leading to altered circadian rhythms	clock mutants help study the role of specific components in circadian rhythms. e.g. full-body Bmal1 knockout mice, lose all circadian rhythms and suffer from health issues, so tissue-specific KO models are often used	misconception: all mutations in clock genes result in complete rhythm loss clarification: some mutations may only alter the period or phase of the rhythm rather than abolishing it entirely; for clock genes with additional functions, those are also affected
clock-controlled gene	a gene whose expression is influenced by the circadian clock but is not part of the clock itself	these genes show circadian patterns of expression but are regulated by clock genes	misconception: clock-regulated genes are components of the clock mechanism clarification: clock-regulated genes are influenced by the circadian clock but do not form part of its core mechanism; the clock ticks normally without them
constant conditions	environmental conditions where an external time cue has been removed, e.g. constant darkness (DD)	these conditions allow the study of an organism's circadian rhythm without external influence; in DD, circadian time CT0−12 corresponds to the subjective day and CT12−24 corresponds to the subjective night	misconception: constant conditions dampen rhythms clarification: constant conditions reveal the organism's internal rhythms, which continue to persist without external cues
coupling/ decoupling	coupling allows the synchronization of circadian rhythms among cellular clocks or between the SCN and peripheral clocks, or enables synchrony between individual cells, or organisms	coupling involves intercellular communication (e.g. via signalling molecules or electrical activity) that ensures alignment of cell rhythms and synchronization; decoupling can occur owing to clock disruptions or pathological states, leading to misaligned rhythms or loss of synchrony	misconception: decoupling always indicates a pathological or negative state clarification: decoupling can be an adaptive response to changing environmental conditions; not all cellular rhythms are perfectly coupled at all times, and coupling strength can vary across tissues and conditions
diel	a rhythm with a 24 h period in the presence of a light–dark cycle	diel is usually used to refer to rhythms that follow 24 h LD cycles, but are not known to be the output of a circadian clock	misconception: diel and diurnal share the same definition clarification: diel refers to a 24 h pattern; diurnal means active predominantly during the light phase
diurnal (nocturnal)	active during the day (or night).	diurnal animals are awake and perform most activities during daylight and nocturnal animals at night; diurnal and nocturnal are extremes at opposite ends of a spectrum with other patterns between	misconception: diurnal animals are completely inactive at night clarification: diurnal animals may have some level of activity during the night, but it is minimal compared with daytime activity (vice versa for nocturnal)
entrainment	the process by which external cues (zeitgebers) align the internal circadian clock with the external environment	entrainment primarily occurs through exposure to light, which is the most powerful zeitgeber for mammals. other zeitgebers include food, temperature, and social interactions; entrainment allows the circadian clock to adapt to changes such as seasonal light–dark cycles	misconception: entrainment is synonymous with synchronization clarification: entrainment specifically involves the process of resetting the internal clock based on external cues, while synchronization can refer to the alignment of internal rhythms with each other or with external cues, which may occur independently of entrainment
free running	the natural rhythm observed under constant conditions, reflecting the organism's internal clock	the free-running period is typically close to 24 h but may be shorter or longer depending on the species.	misconception: free-running rhythms are unpredictable clarification: free-running rhythms are predictable and follow the internal circadian clock, even without external cues
gating	the circadian clock restricts biological processes or responses to particular phases in the 24 h cycle, creating time windows of increased sensitivity or activity	zeitgebers can be subject to gating, and gating can occur to responses to both exogenous and endogenous cues or stimuli (e.g. light, hormones, pathogens); gating explains why functions such as DNA repair or immune responses occur more frequently at certain times of day	misconception: gating is often mistaken for rhythmicity clarification: rhythmicity describes regular cycles, while gating adds a control layer; while gating increases sensitivity at specific times, processes can still happen outside these windows, with reduced efficiency
jetlag	a temporary disruption of the circadian rhythm caused by rapid travel across multiple time zones, leading to a misalignment between the internal clock and the new local time	symptoms of jetlag include sleep disturbances, daytime fatigue, cognitive impairments and digestive issues; recovery time varies based on the number of time zones crossed and the direction of travel (eastward versus westward)	misconception: jetlag is just feeling tired from travel clarification: jetlag involves a comprehensive misalignment of circadian rhythms, affecting multiple physiological processes' eastward travel generally causes more severe jetlag because it requires a phase advance, which is harder to adjust to
light–dark cycle (LD)	a 12 h light/12 h dark cycle is most often used to simulate natural day and night. The duration of the light part of the day is referred to as the photoperiod	the LD cycle is used to synchronize the rhythms of experimental subjects; the durations of light and dark can vary; e.g. 18 h light and 6 h dark simulates a summer photoperiod in temperate zones, and non-24-h cycles (T-cycles) occur when the durations of light + dark are not 24 h.	misconception: rhythms observed under a LD cycle are always driven by a clock clarification: rhythms can be driven by responses to external LD changes, or other environmental signals
masking	a phenomenon where an external factor (e.g. unexpected light exposure) directly affects an organism's behaviour/physiology, temporarily concealing one or more circadian rhythm	masking can obscure the true circadian rhythm(s), making it appear different from what the internal clock dictates	misconception: masking and entrainment are the same clarification: masking prevents the detection of the circadian rhythm, while entrainment is the process of aligning the internal clock to external cues
onset of activity (or rest)	the time point when an organism begins its active phase, often used as a marker of the start of the subjective night in nocturnal animals (vice versa for rest)	for many nocturnal animals under a 12 : 12 LD cycle, the onset of activity typically occurs around lights off. In constant darkness (DD), this is measured by the circadian time (CT12) (vice versa for rest).	misconception: onset of activity is always tied to the external light cue clarification: by constant darkness, activity onset is driven by the internal circadian clock (vice versa for rest)
phase, phase change/shift	a specific point in the rhythm cycle, such as the timing of peak activity; phase advance: a shift in the circadian clock to an earlier time (e.g. waking up earlier); phase delay: a shift to a later time (e.g. going to sleep later); phase shift: a general term for any change in the timing of the circadian rhythm	phase depends on external cues like light, or internal factors; phase shifts are part of entrainment and occur naturally (e.g. during adolescence or aging) or owing to external factors (e.g. light exposure, jetlag); phase advances and delays refer to specific types of shifts; for example, a phase advance would mean the rhythm starts earlier than usual, while a phase delay means it starts later	misconception: phase shifts only indicate a disorder clarification: phase shifts can be normal adaptations or responses to environmental changes, such as adjusting to a new sleep schedule or seasonal changes in daylight or travel across time zones
period	the duration of one complete cycle of a rhythmic process, such as the time from one peak of activity to the next	in circadian rhythms, the period is often close to 24 h but can vary slightly depending on species, individual, and environmental conditions	misconception: circadian period must be exactly 24 h clarification: the circadian period can vary slightly, typically ranging from 23 to 25 h, depending on the species and conditions
subjective day (or night)	the phase of the circadian cycle that corresponds to the rest period for nocturnal animals or the active period for diurnal animals, occurring in the absence of light cues (vice versa for subjective night)	by constant darkness (DD), the subjective day for nocturnal animals corresponds to CT0−12; for diurnal animals, subjective day is the phase when they would normally be active if light cues were present (vice versa for subjective night)	misconception: subjective day is the same as actual day clarification: subjective day is defined by the organism's internal clock and may not align with the external day, especially under constant conditions (vice versa for subjective night)
synchronization	the alignment or coordination of an organism's internal biological rhythms, either within the body or with external environmental cues	essential for maintaining overall physiological coherence and health; disruptions in synchronization can lead to various health problems, including metabolic disorders, immune dysfunction and mood disturbances	misconception: synchronization is often confused with cell cycle synchronization, which refers to the alignment of cell division phases in a population of cells clarification: cell division synchronization may involve timing, but via different mechanisms than circadian synchronization; synchronization should be defined and whether it refers to circadian or cell populations made clear
temperature compensation	a mechanism that ensures the circadian clock maintains its stable period over different operating temperatures	the speed of a clock with temperature-sensitive molecular reactions would vary according to the body temperature of its owner; environmental temperature can act as a Zeitgeber	misconception: temperature has no impact on circadian timekeeping clarification: clocks are only compensated across a biologically relevant temperature range
time-of-day variation	a quantity of a biological system differs according to the time-of-day when it is measured	day (light phase) and night (dark phase) are often compared	misconception: evidence for a circadian clock clarification: clock driven, and non-endogenous rhythms, cause time-of-day variation
zeitgeber	‘time giver'; external cue (e.g. light) that aligns biological rhythms via entrainment.	zeitgebers are used to align rhythms to environmental time, like resetting a watch to the correct time. Zeitgeber time (ZT) usually refers to hours since dawn (or lights on)	misconception: zeitgebers are only light-related cues clarification: other factors such as temperature, social interactions and feeding schedules can also act as zeitgebers

Ethics

This work did not require ethical approval from a human subject or animal welfare committee.

Declaration of AI use

We have not used AI-assisted technologies in creating this article.

Authors’ contributions

R.S.E.: conceptualization, funding acquisition, project administration, writing—original draft, writing—review and editing; A.J.O'D.: data curation, formal analysis, investigation, visualization, writing—review and editing; A.X.Z.: data curation, formal analysis, funding acquisition, investigation, validation, writing—review and editing; S.E.R.: conceptualization, funding acquisition, investigation, project administration, supervision, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed therein.

Conflict of interest declaration

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Funding

R.S.E. is supported by Royal Society—Wellcome Sir Henry Dale Fellowship (208790/Z/17/Z). A.X.Z. is supported by the MRC Career Development Award (MR/X020843/1). S.E.R. and A.J.O'D. are supported by Wellcome (202769/Z/16/Z and 227904/Z/23/Z).