Capture-Mark-Recapture (CMR) data are commonly used to estimate ecological variables such as abundance, survival probability, or transition rates from one state to another (e.g. from juvenile to adult, or migration from one site to another). Many studies have shown how estimations can be affected by neglecting one aspect of the population under study (e.g. the heterogeneity in survival between individuals) or one limit of the methodology itself (e.g. the fact that observers might not detect an individual although it is still alive). Strikingly, very few studies have yet assessed the robustness of one fundamental assumption of all CMR-based inferences: marks are supposed definitive and immutable. If they are not, how are estimations affected? Addressing this issue is the main goal of the paper by Touzalin et al. (2023), and they did a very nice work. But, because the answer is not that simple, it also calls for further investigations.

When and why would mark loss bias estimation? In at least two situations. First, when estimating survival rates: if an individual loses its mark, it will be considered as dead, hence death rates will be overestimated. Second, more subtly, when estimating transition rates: if one individual loses its mark at the specific moment where its state changes, then a transition will be missed in data. The history of the marked individual would then be split into two independent CMR sequences as if there were two different individuals, including one which died.

Touzalin et al. (2023) thoroughly studied these two situations by estimating ecological parameters on 1) well-thought simulated datasets, that cover a large range of possible situations inspired from a nice compilation of hundreds of estimations from fish and bats studies, and 2) on their own bats dataset, for which they had various sources of information about mark losses, i.e. different mark types on the same individuals, including mark based on genotypes, and marks found on the soil in the place where bats lived. Their main findings from the simulated datasets are that there is a general trend for underestimation of survival and transition rates if mark loss is not accounting for in the model, as it would be intuitively expected. However, they also showed from the bats dataset that biases do not show any obvious general trend, suggesting complex interactions between different ecological processes and/or with the estimation procedure itself.

The results by Touzalin et al. (2023) strongly suggest that mark loss should systematically be included in models estimating parameters from CMR data. In addition to adapt the inferential models, the authors also recommend considering either a double marking, or even a single but ‘permanent’ mark such as one based on the genotypes. However, the potential gain of a double marking or of the use of genotypes is still to be evaluated both in theory and practice, and it seems to be not that obvious at first sight. First because double marking can be costly for experimenters but also for the marked animals, especially as several studies showed that marks can significantly affect survival or recapture rates. Second because multiple sources of errors can affect genotyping, which would result in wrong individual assignations especially in populations with low genetic diversity or high inbreeding, or no individual assignation at all, which would increase the occurrence of missing data in CMR datasets. Touzalin et al. (2023) supposed in their paper that there were no genotyping errors, but one can doubt it to be true in most situations. They have now important and interesting other issues to address.

References

Frédéric Touzalin, Eric J. Petit, Emmanuelle Cam, Claire Stagier, Emma C. Teeling, Sébastien J. Puechmaille (2023) Mark loss can strongly bias demographic rates in multi-state models: a case study with simulated and empirical datasets. BioRxiv, ver. 3 peer-reviewed and recommended by Peer Community in Ecology. https://doi.org/10.1101/2022.03.25.485763

The ability to successfully sequence trace quantities of environmental DNA (eDNA) has provided unprecedented opportunities to use genetic analyses to elucidate animal ecology, behavior, and population structure without affecting the behavior, fitness, or welfare of the animal sampled. Hair associated with an animal track in the snow, the shed exoskeleton of an insect, or a swab of animal scat are all examples of non-invasive methods to collect eDNA. Despite the seemingly uncomplicated definition of "non-invasive" as proposed by Taberlet et al. [1], Lefort et al. [2] highlight that its appropriate application to sampling methods in practice is not so straightforward. For example, collecting scat left behind on the forest floor by a mammal could be invasive if feces is used by that species to mark territorial boundaries. Other collection strategies such as baited DNA traps to collect hair, capturing and handling an individual to swab or stimulate emission of a body fluid, or removal of a presumed non essential body part like a feather, fish scale, or even a leg from an insect are often described as "non-invasive" sampling methods. However, such methods cannot be considered truly non-invasive. At a minimum, attracting or capturing and handling an animal to obtain a DNA sample interrupts its normal behavioral routine, but additionally can cause both acute and long-lasting physiological and behavioral stress responses and other effects. Even invertebrates exhibit long-term hypersensitization after an injury, which manifests as heightened vigilance and enhanced escape responses [3-5].
Through an extensive analysis of 380 papers published from 2013-2018, Lefort et al. [2] document the widespread misapplication of the term "non-invasive" to methods used to sample DNA. An astonishing 58% of these papers employed the term incorrectly. A big part of the problem is that "non-invasive" is usually used by authors in the medical or veterinary sense of not breaking the skin or entering the body [6], rather than in the broader, ecological sense of Taberlet et al. [1]. The authors argue that correct use of the term matters, because it may lead naive readers – one can imagine students, policy makers, and the general public – to incorrectly assume a particular method is safe to use in a situation where disturbing the animal could affect experimental results or raise animal welfare concerns. Such assumptions can affect experimental design, as well as interpretations of one's own or others' data.
The importance of the Lefort et al. [2] paper lies in part on the authors' call for the research community to be much more careful when applying the term "non-invasive" to methods of DNA sampling. This call cannot be shrugged off as a minor problem in a few papers – as their literature review demonstrates, "non-invasive" is being applied incorrectly more often than not. The authors recognize that not all DNA sampling must be non-invasive to be useful or ethical. Examples include taking samples for DNA extraction from museum specimens, or opportunistically from carcasses of animals hunted either legally or seized by authorities from poachers. In many cases, there may be no viable non-invasive method to obtain DNA, but a researcher strives to collect samples using methods that, although they may involve taking a sample directly from the animal's body, do not disrupt, or only slightly disrupt behavior, fitness, or welfare of the animal. Thus, the other important contribution by Lefort et al. [2] is to propose the terms "non-disruptive" and "minimally-disruptive" to describe such sampling methods, which are not strictly non-invasive. While gray areas undoubtedly remain, as acknowledged by the authors, answering the call for correct use of "non-invasive" and applying the proposed new terms for certain types of invasive sampling with a focus on level of disruption, will go a long way in limiting misconceptions and misinterpretations caused by the current confusion in terminology.

References

[1] Taberlet P., Waits L. P. and Luikart G. 1999. Noninvasive genetic sampling: look before you leap. Trends Ecol. Evol. 14: 323-327. doi: 10.1016/S0169-5347(99)01637-7
[2] Lefort M.-C., Cruickshank R. H., Descovich K., Adams N. J., Barun A., Emami-Khoyi A., Ridden J., Smith V. R., Sprague R., Waterhouse B. R. and Boyer S. 2019. Blood, sweat and tears: a review of non-invasive DNA sampling. bioRxiv, 385120, ver. 4 peer-reviewed and recommended by PCI Ecology. doi: 10.1101/385120
[3] Khuong T. M., Wang Q.-P., Manion J., Oyston L. J., Lau M.-T., Towler H., Lin Y. Q. and Neely G. G. 2019. Nerve injury drives a heightened state of vigilance and neuropathic sensitization in Drosophila. Science Advances 5: eaaw4099. doi: 10.1126/sciadv.aaw4099
[4] Crook, R. J., Hanlon, R. T. and Walters, E. T. 2013. Squid have nociceptors that display widespread long-term sensitization and spontaneous activity after bodily injury. Journal of Neuroscience, 33(24), 10021-10026. doi: 10.1523/JNEUROSCI.0646-13.2013
[5] Walters E. T. 2018. Nociceptive biology of molluscs and arthropods: evolutionary clues about functions and mechanisms potentially related to pain. Frontiers in Physiololgy 9: doi: 10.3389/fphys.2018.01049
[6] Garshelis, D. L. 2006. On the allure of noninvasive genetic sampling-putting a face to the name. Ursus 17: 109-123. doi: 10.2192/1537-6176(2006)17[109:OTAONG]2.0.CO;2

Alternative stable states describe ecosystems that can persist in more than one configuration. An ecosystem can shift between stable states following some form of perturbation. There has been much work on predicting when ecosystems will shift between stable states, but less work on why some ecosystems are able to exist in alternative stable states in the first place. The paper by Ardichvili, Loeuille, and Dakos (2022) addresses this question using a simple model of a shallow lake. Their model is based on a trade-off between access to light and nutrient availability in the water column, two essential resources for the macrophytes they model. They then identify conditions when the ancestral macrophyte will diversify resulting in macrophyte species living at new depths within the lake. The authors find a range of conditions where alternative stable states can evolve, but the range is narrow. Nonetheless, their model suggests that for alternative stable states to exist, one requirement is for there to be asymmetric competition between competing species, with one species being a better competitor on one limiting resource, with the other being a better competitor on a second limiting resource. 

These results are interesting and add to growing literature on how asymmetric competition can aid species coexistence. Asymmetric competition may be widespread in nature, with closely related species often being superior competitors on different resources. Incorporating asymmetric competition, and its evolution, into models does complicate theoretical investigations, but Ardichvili, Loeuille, and Dakos’ paper elegantly shows how substantial progress can be made with a model that is still (relatively) simple.

References 

Ardichvili A, Loeuille N, Dakos V (2022) Evolutionary emergence of alternative stable states in shallow lakes. bioRxiv, 2022.02.23.481597, ver. 3 peer-reviewed and recommended by Peer Community in Ecology. https://doi.org/10.1101/2022.02.23.481597

Trophic interactions are at the heart of community ecology. Herbivores consume plants, predators consume herbivores, and pathogens and parasites infect, and sometimes kill, individuals of all species in a food web. Given the ubiquity of trophic interactions, it is no surprise that ecologists and evolutionary biologists strive to accurately characterize them. 

The outcome of an interaction between individuals of different species depends upon numerous factors such as the age, sex, and even phenotype of the individuals involved and the environment in which they are in. Despite this complexity, biologists often simplify an interaction down to a single number, an interaction coefficient that describes the average outcome of interactions between members of the populations of the species. Models of interacting species tend to be very simple, and interaction coefficients are often estimated from time series of population sizes of interacting species. Although biologists have long known that this approach is often approximate and sometimes unsatisfactory, work on estimating interaction strengths in more complex scenarios, and using ecological data beyond estimates of abundance, is still in its infancy. 

In their paper, Matthieu Paquet and Frederic Barraquand (2023)​ develop a demographic model of a predator and its prey. They then simulate demographic datasets that are typical of those collected by ecologists and use integrated population modelling to explore whether they can accurately retrieve the values interaction coefficients included in their model. They show that they can with good precision and accuracy. The work takes an important step in showing that accurate interaction coefficients can be estimated from the types of individual-based data that field biologists routinely collect, and it paves for future work in this area.

As if often the case with exciting papers such as this, the work opens up a number of other avenues for future research. What happens as we move from demographic models of two species interacting such as those used by Paquet and Barraquand​ to more realistic scenarios including multiple species? How robust is the approach to incorrectly specified process or observation models, core components of integrated population modelling that require detailed knowledge of the system under study? 

Integrated population models have become a powerful and widely used tool in single-species population ecology. It is high time the techniques are extended to community ecology, and this work takes an important step in showing that this should and can be done. I would hope the paper is widely read and cited.

References

Paquet, M., & Barraquand, F. (2023). Assessing species interactions using integrated predator-prey models. EcoEvoRxiv, ver. 2 peer-reviewed and recommended by Peer Community in Ecology. https://doi.org/10.32942/X2RC7W

Tick-borne diseases are an important burden on public health all over the globe, making accurate forecasts of tick population a key ingredient in a successful public health strategy. Over long time scales, tick populations can undergo complex dynamics, as they are sensitive to many non-linear effects due to the complex relationships between ticks and the relevant (numerical) features of their environment.

But luckily, capturing complex non-linear responses is a task that machine learning thrives on. In this contribution, Manley et al. (2023) explore the use of Gradient Boosted Trees to predict the distribution (presence/absence) and abundance of ticks across New York state.

This is an interesting modelling challenge in and of itself, as it looks at the same ecological question as an instance of a classification problem (presence/absence) or of a regression problem (abundance). In using the same family of algorithm for both, Manley et al. (2023) provide an interesting showcase of the versatility of these techniques. But their article goes one step further, by setting up a multi-class categorical model that estimates jointly the presence and abundance of a population. I found this part of the article particularly elegant, as it provides an intermediate modelling strategy, in between having two disconnected models for distribution and abundance, and having nested models where abundance is only predicted for the present class (see e.g. Boulangeat et al., 2012, for a great description of the later).

One thing that Manley et al. (2023) should be commended for is their focus on opening up the black box of machine learning techniques. I have never believed that ML models are more inherently opaque than other families of models, but the focus in this article on explainable machine learning shows how these models might, in fact, bring us closer to a phenomenological understanding of the mechanisms underpinning our observations.

There is also an interesting discussion in this article, on the rate of false negatives in the different models that are being benchmarked. Although model selection often comes down to optimizing the overall quality of the confusion matrix (for distribution models, anyway), depending on the type of information we seek to extract from the model, not all types of errors are created equal. If the purpose of the model is to guide actions to control vectors of human pathogens, a false negative (predicting that the vector is absent at a site where it is actually present) is a potentially more damaging outcome, as it can lead to the vector population (and therefore, potentially, transmission) increasing unchecked.

Boulangeat I, Gravel D, Thuiller W. Accounting for dispersal and biotic interactions to disentangle the drivers of species distributions and their abundances: The role of dispersal and biotic interactions in explaining species distributions and abundances. Ecol Lett. 2012;15: 584-593.
https://doi.org/10.1111/j.1461-0248.2012.01772.x

Manley W, Tran T, Prusinski M, Brisson D. (2023) Modeling tick populations: An ecological test case for gradient boosted trees. bioRxiv, 2023.03.13.532443, ver. 3 peer-reviewed and recommended by Peer Community in Ecology. https://doi.org/10.1101/2023.03.13.532443