Automated location invariant animal detection in camera trap images using publicly available data sources

1. A time‐consuming challenge faced by camera trap practitioners is the extraction of meaningful data from images to inform ecological management. An increasingly popular solution is automated image classification software. However, most solutions are not sufficiently robust to be deployed on a large scale due to lack of location invariance when transferring models between sites. This prevents optimal use of ecological data resulting in significant expenditure of time and resources to annotate and retrain deep learning models.
2. We present a method ecologists can use to develop optimized location invariant camera trap object detectors by (a) evaluating publicly available image datasets characterized by high intradataset variability in training deep learning models for camera trap object detection and (b) using small subsets of camera trap images to optimize models for high accuracy domain‐specific applications.
3. We collected and annotated three datasets of images of striped hyena, rhinoceros, and pigs, from the image‐sharing websites FlickR and iNaturalist (FiN), to train three object detection models. We compared the performance of these models to that of three models trained on the Wildlife Conservation Society and Camera CATalogue datasets, when tested on out‐of‐sample Snapshot Serengeti datasets. We then increased FiN model robustness by infusing small subsets of camera trap images into training.
4. In all experiments, the mean Average Precision (mAP) of the FiN trained models was significantly higher (82.33%–88.59%) than that achieved by the models trained only on camera trap datasets (38.5%–66.74%). Infusion further improved mAP by 1.78%–32.08%.
5. Ecologists can use FiN images for training deep learning object detection solutions for camera trap image processing to develop location invariant, robust, out‐of‐the‐box software. Models can be further optimized by infusion of 5%–10% camera trap images into training data. This would allow AI technologies to be deployed on a large scale in ecological applications. Datasets and code related to this study are open source and available on this repository: https://doi.org/10.5061/dryad.1c59zw3tx.

     

Large‐scale survey and database of high affinity ligands for peptide recognition modules

Many proteins involved in signal transduction contain peptide recognition modules (PRMs) that recognize short linear motifs (SLiMs) within their interaction partners. Here, we used large‐scale peptide‐phage display methods to derive optimal ligands for 163 unique PRMs representing 79 distinct structural families. We combined the new data with previous data that we collected for the large SH3, PDZ, and WW domain families to assemble a database containing 7,984 unique peptide ligands for 500 PRMs representing 82 structural families. For 74 PRMs, we acquired enough new data to map the specificity profiles in detail and derived position weight matrices and binding specificity logos based on multiple peptide ligands. These analyses showed that optimal peptide ligands resembled peptides observed in existing structures of PRM‐ligand complexes, indicating that a large majority of the phage‐derived peptides are likely to target natural peptide‐binding sites and could thus act as inhibitors of natural protein–protein interactions. The complete dataset has been assembled in an online database (http://www.prm‐db.org) that will enable many structural, functional, and biological studies of PRMs and SLiMs.     

Membranome 3.0: Database of single‐pass membrane proteins with AlphaFold models

The Membranome database provides comprehensive structural information on single‐pass (i.e., bitopic) membrane proteins from six evolutionarily distant organisms, including protein–protein interactions, complexes, mutations, experimental structures, and models of transmembrane α‐helical dimers. We present a new version of this database, Membranome 3.0, which was significantly updated by revising the set of 5,758 bitopic proteins and incorporating models generated by AlphaFold 2 in the database. The AlphaFold models were parsed into structural domains located at the different membrane sides, modified to exclude low‐confidence unstructured terminal regions and signal sequences, validated through comparison with available experimental structures, and positioned with respect to membrane boundaries. Membranome 3.0 was re‐developed to facilitate visualization and comparative analysis of multiple 3D structures of proteins that belong to a specified family, complex, biological pathway, or membrane type. New tools for advanced search and analysis of proteins, their interactions, complexes, and mutations were included. The database is freely accessible at https://membranome.org.     

The Data Governance Act and the EU's move towards facilitating data sharing

The implementation of the EU General Data Protection Regulation (GDPR) has had significant impacts on biomedical research, often complicating data sharing among researchers. The recently announced proposal for a new EU Data Governance Act is a promising step towards facilitating data sharing, if it can interplay well with the GDPR.Subject Categories: S&S: Ethics

The EU General Data Protection Regulation (GDPR) has affected biomedical research, often complicating data sharing. The recently announced proposal for a new EU Data Governance Act, is a promising step towards facilitating data sharing.

In an attempt to improve and increase data sharing in the EU and to optimize the re‐use of personal and non‐personal data, the European Commission has recently announced the proposal for a new EU Data Governance Act (https://ec.europa.eu/digital‐single‐market/en/news/proposal‐regulation‐european‐data‐governance‐data‐governance‐act). If approved, it will enable the creation and regulation of “secure spaces” where various types of data, including health data, can be shared and re‐used for both commercial and altruistic purposes, including scientific research. The Data Governance Act, within the framework of a European Strategy for Data, (https://ec.europa.eu/info/sites/info/files/communication‐european‐strategy‐dat‐19feb2020_en.pdf), would address some of the shortcomings and drawbacks of the current regulatory framework which holds back sharing and re‐using data for biomedical research purposes.While the proposed Act would apply to all types of personal and non‐personal data, the increasing demand for sharing health data has most likely been a major rationale for this new legislation of data governance. Notably, sharing health and genetic data for scientific research entails an extra layer of complexity, owing to concerns over data protection and privacy when sharing sensitive personal data. Vice versa, there are also concerns in the scientific community over the negative impact of regulatory restrictions on sharing health data in data‐driven biomedical research. The pressing question here is how far the EU’s proposed legislative and policy framework can offset either concerns?     

Biodiversity 2030: a road paved with good intentions: The new EU Commission's biodiversity Strategy risks to remain an empty husk without proper implementation

The EU''s Biodiversity Strategy for 2030 makes great promises about halting the decline of biodiversity but it offers little in terms of implementation. Subject Categories: S&S: Economics & Business, Ecology, S&S: Ethics

Earth is teeming with a stunning variety of life forms. Despite hundreds of years of exploration and taxonomic research, and with 1.2 million species classified, we still have no clear picture of the real extent of global biodiversity, with estimates ranging from 3 to 100 million species. A highly quoted—although not universally accepted—study predicted some 8.7 million species, of which about 2.2 million are marine (Mora et al, 2011). Although nearly any niche on the surface of Earth has been colonized by life, species richness is all but evenly distributed. A large share of the known species is concentrated in relatively small areas, especially in the tropics (Fig 1). Ultimately, it is the network of the interactions among life forms and the physical environment that make up the global ecosystem we call biosphere and that supports life itself.Open in a separate windowFigure 1Biological hotspots of the worldA total of 36 currently recognized hotspots make up < 3% of the planet''s land area but harbor half of the world''s endemic plant species and 42% of all terrestrial vertebrates. Overall, hotspots have lost more than 80% of their original extension. Credit: Richard J. Weller, Claire Hoch, and Chieh Huang, 2017, Atlas for the End of the World, http://atlas‐for‐the‐end‐of‐the‐world.com/. Reproduced with permission.Driven by a range of complex and interwoven causes–such as changes in land and sea use, habitat destruction, overexploitation of organisms, climate change, pollution, and invasive species–biodiversity is declining at an alarming pace. A report by the Intergovernmental Science‐Policy Platform on Biodiversity and Ecosystem Services (IPBES) issued a clear warning: “An average of around 25 per cent of species in assessed animal and plant groups are threatened, suggesting that around 1 million species already face extinction, many within decades, unless action is taken to reduce the intensity of drivers of biodiversity loss. Without such action, there will be a further acceleration in the global rate of species extinction, which is already at least tens to hundreds of times higher than it has averaged over the past 10 million years” (IPBES, 2019) (Fig 2). Although focused on a smaller set of organisms, a more recent assessment by WWF has reached similar conclusions. Their Living Planet Index, that tracks the abundance of thousands of populations of mammals, birds, fish, reptiles, and amphibians around the world, shows a stark decline in monitored populations (WWF, 2020). As expected, the trend of biodiversity decline is not homogeneous with tropical areas paying a disproportionately high price, mostly because of unrestrained deforestation and exploitation of natural resources.Open in a separate windowFigure 2The global, rapid decline of biodiversity(A) Percentage of species threatened with extinction in taxonomic groups that have been assessed comprehensively, or through a “sampled” approach, or for which selected subsets have been assessed by the IUCN Red List of Threatened Species. Groups are ordered according to the best estimate, assuming that data‐deficient species are as threatened as non‐data deficient species. (B) Extinctions since 1500 for vertebrate groups. (C) Red List Index of species survival for taxonomic groups that have been assessed for the IUCN Red List at least twice. A value of 1 is equivalent to all species being categorized as Least Concern; a value of zero is equivalent to all species being classified as Extinct. Data for all panels from www.iucnredlist.org. Reproduced from (IPBES, 2019), with permission.

Driven by a range of complex and interwoven causes […] biodiversity is declining at an alarming pace.

Against this dire background, the EU has drafted a Biodiversity Strategy 2030, an ambitious framework aimed to tackling the key reasons behind biodiversity loss. The plan hinges around a few main elements, such as the establishment of protected areas for at least 30% of Europe''s lands and seas (Fig 3); a significant increase of biodiversity‐rich landscape features on agricultural land by establishing buffer zones like hedges and fallow fields; halting and reversing the decline of pollinators; and planting 3 billion trees by 2030 (https://ec.europa.eu/info/strategy/priorities‐2019‐2024/european‐green‐deal/actions‐being‐taken‐eu/eu‐biodiversity‐strategy‐2030_en). The budget for implementing these measures was set at €20 billion per year.Open in a separate windowFigure 3Natura 2000, the EU''s network of protected areasIn 2019, 18% of land in the EU was protected as Natura 2000, with the lowest share of protected land in Denmark (8%) and the highest in Slovenia (38%). In 2019, the largest national network of terrestrial Natura 2000 sites was located in Spain, covering 138,111 km², followed by France (70,875 km²) and Poland (61,168 km²). Reproduced from Eurostat: https://ec.europa.eu/eurostat/statistics‐explained/index.php?title=Main_Page “Nature is vital for our physical and mental wellbeing, it filters our air and water, it regulates the climate and it pollinates our crops. But we are acting as if it didn''t matter, and losing it at an unprecedented rate”, said Virginijus Sinkevičius, Commissioner for the Environment, Oceans and Fisheries, at the press launch of the new EU action (https://ec.europa.eu/commission/presscorner/detail/en/ip_20_884). “This new Biodiversity Strategy builds on what has worked in the past, and adds new tools that will set us on a path to true sustainability, with benefits for all. The EU''s aim is to protect and restore nature, to contribute to economic recovery from the current crisis, and to lead the way for an ambitious global framework to protect biodiversity around the planet”.Environmental groups and other stakeholders have welcomed the EU''s pledge in principle. “This is a unique opportunity to shape a new society in harmony with nature”, applauded Wetlands International. “We must not forget that the biodiversity and climate crisis is a much bigger and persistent challenge for humanity than COVID‐19”, (https://europe.wetlands.org/news/welcoming‐the‐eu‐biodiversity‐strategy‐for‐2030/). EuroNatur, a foundation focused on conservation, stated that the goals set out by the new strategy provide a strong basis for improving the state of nature in the EU (www.euronatur.org).Alongside the voices of praise, however, many have expressed concerns that the strategy could turn into a little more than a wish list. “The big issue of the strategy is that while setting a goal for financial funds, the EU does not specify where the money is supposed to come from. It only says it should include ‘EU funds and national and private funding’”, commented the European Wilderness Society, an environmental advocacy non‐profit organization headquartered in Tamsweg, Austria. “Goals are important, but do not create change without an organized and sustainable implementation. It''s a good and ambitious document, but what is also obvious is the lack of strategy of how to implement it, and a lack of discussion of why previous documents of this type failed” (https://wilderness‐society.org/ambitious‐eu‐biodiversity‐strategy‐2030/).

Alongside the voices of praise, however, many have expressed concerns that the strategy could turn into a little more than a wish list.

The Institute for European Environmental Policy (IEEP) is on the same page. The sustainability think‐tank based in Brussels and London noted that the outgoing EU 2020 biodiversity strategy showed major implementation problems, especially because of lack of engagement at national level and of ad hoc legislation supporting the meeting of key targets. Therefore, “[it] can be argued that a legally binding approach to the biodiversity governance framework is urgently needed unless Member States and other key stakeholders can show greater intrinsic ownership to deliver on agreed objectives”, (https://ieep.eu/news/first‐impressions‐of‐the‐eu‐biodiversity‐strategy‐to‐2030). In addition, IEEP remarked that money is an issue, since the €20 billion figure appears more as an estimate than a certified obligation.“The intentions of the Commission are good and the strategy contains a number of measures and targets that can really make a difference. However, implementation depends critically on the member states and experiences with the Common Agricultural Policy the past decade or so have taught us that many of them are more interested in short‐term economic objectives than in safeguarding the natural wealth of their country for future generations”, commented David Kleijn, an ecologist and nature conservation expert at the Wageningen University, the Netherlands. “I think it is important that we now have an ambitious Biodiversity Strategy but at the same time I have little hope that we will be able to achieve its objectives”.

I think it is important that we now have an ambitious Biodiversity Strategy but at the same time I have little hope that we will be able to achieve its objectives.

There is further criticism against specific measures, such as the proposal of planting 3 billion trees. “To have lots of trees planted in an area does not necessarily translate into an increase of biodiversity. Biodiverse ecosystems are the result of million years of complex multi‐species interactions and evolutionary processes, which are not as easy to restore”, explained plant ecologist Susana Gómez‐González, from the University of Cádiz, Spain. Planting a large number of trees is a too simplistic approach for saving European forests from the combined effects of excessive anthropic pressure and climate change, and could even have detrimental effects (see Box 1). More emphasis should be placed instead in reducing tree harvesting in sensitive areas and in promoting natural forest renewal processes (Gómez‐González et al, 2020). “For a biodiversity strategy, increasing the number of trees, or even increasing the forest area, should not be an objective; priority should be given to the conservation and restoration of natural ecosystems, forests and non‐forests”, Gómez‐González said.In other cases, it could be difficult, if not impossible, to reach some of the goals because of lack of information. For example, one of the roadmap''s targets is to restore at least 25,000 km of Europe''s rivers back to free‐flowing state. However, the number of barriers dispersed along European rivers will probably prevent even getting close to the mark. An international research team has collected detailed information on existing instream barriers for 147 rivers in 36 European countries, coming up with the impressive figure of over 1.2 million obstacles that inevitably impact on river ecosystems, affecting the transport and dispersion of aquatic organisms, nutrients, and sediments (Belletti et al, 2020). Existing inventories mainly focused on dams and other large barriers, while, in fact, a large number of artificial structures are much smaller, such like weirs, locks, ramps, and fords. As a result, river fragmentation has been largely underestimated, and the models used to plan flow restoration might be seriously flawed. “To avoid ‘death by a thousand cuts’, a paradigm shift is necessary: to recognize that although large dams may draw most of the attention, it is the small barriers that collectively do most of the damage. Small is not beautiful”, concluded the authors (Belletti et al, 2020).

Box 1: Why many trees don''t (always) make a forestForests are cathedrals of biodiversity. They host by far the largest number of species on land, which provide food and essential resources for hundreds of millions of people worldwide. However, forests are disappearing and degrading at an alarming pace. The loss of these crucial ecosystems has given new impulses to a variety of projects aimed at stopping this devastation and possibly reversing the trend.Once it is gone, can you rebuild a forest? Many believe the answer is yes, and the obvious solution is to plant trees. Several countries have thus launched massive tree‐planting programs, notably India and Ethiopia, where 350 million trees have been planted in single day (https://www.unenvironment.org/news‐and‐stories/story/ethiopia‐plants‐over‐350‐million‐trees‐day‐setting‐new‐world‐record). The World Economic Forum has set up its own One Trillion Tree initiative (https://www.1t.org/) “to conserve, restore, and grow one trillion trees by 2030”. Launched in January last year at Davos, 1t.org was conceived as a platform for governments, companies and NGOs/civil society groups to support the UN Decade on Ecosystem Restoration (2021–2030). The initiative has been christened by renowned naturalist Jane Goodall, who commented: “1t.org offers innovative technologies which will serve to connect tens of thousands of small and large groups around the world that are engaged in tree planting and forest restoration”, (https://www.weforum.org/agenda/2020/01/one‐trillion‐trees‐world‐economic‐forum‐launches‐plan‐to‐help‐nature‐and‐the‐climate/).However, things are way more complicated than they appear: large‐scale tree planting schemes are rarely a viable solution and can even be harmful. “[A] large body of literature shows that even the best planned restoration projects rarely fully recover the biodiversity of intact forests, owing to a lack of sources of forest‐dependent flora and fauna in deforested landscapes, as well as degraded abiotic conditions resulting from anthropogenic activities”, commented Karen Holl from the University of Caliornia, Santa Cruz, and Pedro Brancalion from the University of São Paulo (Holl & Brancalion, 2020). A common problem of tree plantations, for example, is the low survival rate of seedlings, mostly because the wrong tree species are selected and due to poor maintenance after planting. Moreover, grasslands and savannas, which are often targeted for establishing new forests, are themselves treasure troves of biodiversity. Ending indiscriminate deforestation, improving the protection of existing forests, and promoting their restoration would therefore be a more efficient strategy to preserve biodiversity in the shorter term. If tree planting is indeed necessary, it should be well planned by selecting the right areas for reforestation, using suitable tree species that can maximize biodiversity, and involving local populations to maintain the plantations, Holl and Brancalion argue (Holl & Brancalion, 2020).

…even the best planned restoration projects rarely fully recover the biodiversity of intact forests, owing to a lack of sources of forest‐dependent flora and fauna in deforested landscapes…

The health of soil, where a high proportion of biodiversity is hosted, is another problem the new strategy should address in a more focused manner. “In my opinion, the EU Biodiversity Strategy is already a leap forward in terms of policy interest in soils in general and in soil biodiversity in particular. Compared with other nations/regions of the world, Europe is by far in the forefront regarding this issue”, commented Carlos António Guerra at the German Centre for Integrative Biodiversity Research (iDiv) in Leipzig, Germany, and Co‐leader of the Global Soil Biodiversity Observation Network (https://geobon.org/bons/thematic‐bon/soil‐bon/). “Nevertheless, the connection between soil biodiversity and ecological functions needs further commitments. Soils allow for horizontal integration of several policy agendas, from climate to agriculture and, very importantly, nature conservation. This is not explicit in the EU Biodiversity Strategy in regard to soils”. It remains to be seen if EU restoration plan will emphasize soil biodiversity, or consider it as a mere side effect of other initiatives, Guerra added. “A soil nature conservation plan should be proposed”, he said. “Only such a plan, that implies that current and future protected areas have to consider, describe and protect their soil biodiversity would make a significant push to help protect such a valuable resource”.More generally, research shows that the current paradigm of protection must be shifted to prevent further losses to biodiversity. In fact, an analysis of LIFE projects—a cornerstone of EU nature protection—found that conservation efforts are extremely polarized and strongly taxonomically biased (Mammola et al, 2020). From 1992 to 2018, investment in vertebrates was sixfold higher than that for invertebrates, with birds and mammals alone accounting for 72% of the targeted species and 75% of the total budget. In relative terms, investment per species for vertebrates has been 468 times higher than for invertebrates (Fig 4). There is no sound scientific reasoning behind this uneven conservation attention, but just popularity. “[T]he species covered by a greater number of LIFE projects were also those which attracted the most interest online, suggesting that conservation in the EU is largely driven by species charisma, rather than objective features”, the researchers wrote (Mammola et al, 2020).Open in a separate windowFigure 4Taxonomic bias in EU fauna protection effortsBreakdown of the number of projects (A) and budget allocation (B) across main animal groups covered by the LIFE projects (n = 835). (C) The most covered 30 species of vertebrates (out of 410) and invertebrates (out of 78) in the LIFE projects analyzed (n = 835). The vertical bar represents monetary investment and the blue scatter line the number of LIFE projects devoted to each species. Reproduced from (Mammola et al, 2020), with permission.     

Reply to Bronstein and Vinogradov

The response by the author. Subject Categories: S&S: Economics & Business, S&S: Ethics

I thank Michael Bronstein and Sophia Vinogradov for their interest and comments. I would like to respond to a few of their points.First, I agree with the authors that empirical studies should be conducted to validate any approaches to prevent the spread of misinformation before their implementation. Nonetheless, I think that the ideas I have proposed may be worth further discussion and inspire empirical studies to test their effectiveness.Second, the authors warn that informing about the imperfections of scientific research may undermine trust in science and scientists, which could result in higher vulnerability to online health misinformation (Roozenbeek et al, 2020; Bronstein & Vinogradov, 2021). I believe that transparency about limitations and problems in research does not necessarily have to diminish trust in science and scientists. On the contrary, as Veit et al put it, “such honesty… is a prerequisite for maintaining a trusting relationship between medical institutions (and practitioners) and the public” (Veit et al, 2021). Importantly, to give an honest picture of scientific research, information about its limitations should be put in adequate context. In particular, the public also should be aware that “good science” is being done by many researchers; we do have solid evidence of effectiveness of many medical interventions; and efforts are being taken to address the problems related to quality of research.Third, Bronstein and Vinogradov suggest that false and dangerous information should be censored. I agree with the authors that “[c]ensorship can prevent individuals from being exposed to false and potentially dangerous ideas” (Bronstein & Vinogradov, 2021). I also recognize that some information is false beyond any doubt and its spread may be harmful. What I am concerned about are, among others, the challenges related to defining what is dangerous and false information and limiting censorship only to this kind of information. For example, on what sources should decisions to censor be based and who should make such decisions? Anyone, whether an individual or an organization, with a responsibility to censor information will likely not only be prone to mistakes, but also to abuses of power to foster their interests. Do the benefits we want to achieve by censorship outweigh the potential risks?Fourth, we need rigorous empirical studies examining the actual impact of medical misinformation. What exactly are the harms we try to protect against and what is their scale? This information is necessary to choose proportionte and effective measures to reduce the harms. Bronstein and Vinogradov give an example of a harm which may be caused by misinformation—an increase in methanol poisoning in Iran. Yet, as noticed by the authors, misinformation is not the sole factor in this case; there are also cultural and other contexts (Arasteh et al, 2020; Bronstein & Vinogradov, 2021). Importantly, the methods of studies exploring the effects of misinformation should be carefully elaborated, especially when study participants are asked to self‐report. A recent study suggests that some claims about the prevalence of dangerous behaviors, such as drinking bleach, which may have been caused by misinformation are largely exaggerated due to the presence of problematic respondents in surveys (preprint: Litman et al, 2021).Last but not least, I would like to call attention to the importance of how veracity of information is determined in empirical studies on misinformation. For example, in a study of Roozenbeek et al, cited by Bronstein and Vinogradov, the World Health Organization (WHO) was used as reliable source of information, which raises questions. For instance, Roozenbeek et al (2020) used a statement “the coronavirus was bioengineered in a military lab in Wuhan” as an example of false information, relying on the judgment of the WHO found on its “mythbusters” website (Roozenbeek et al, 2020). Yet, is there a solid evidence to claim that this statement is false? At present, at least some scientists declare that we cannot rule out that the virus was genetically manipulated in a laboratory (Relman, 2020; Segreto & Deigin, 2020). Interestingly, the WHO also no longer excludes such a possibility and has launched an investigation on this issue (https://www.who.int/health‐topics/coronavirus/origins‐of‐the‐virus, https://www.who.int/emergencies/diseases/novel‐coronavirus‐2019/media‐resources/science‐in‐5/episode‐21‐‐‐covid‐19‐‐‐origins‐of‐the‐sars‐cov‐2‐virus); the information about the laboratory origin of the virus being false is no longer present on the WHO “mythbusters” website (https://www.who.int/emergencies/diseases/novel‐coronavirus‐2019/advice‐for‐public/myth‐busters). Against this backdrop, some results of the study by Roozenbeek et al (2020) seem misleading. In particular, the perception of the reliability of the statement about bioengineered virus by study participants in Roozenbeek et al (2020) does not reflect the susceptibility to misinformation, as intended by the researchers, but rather how the respondents perceive reliability of uncertain information.I hope that discussion and research on these and related issues will continue.     

Predicting insect outbreaks using machine learning: A mountain pine beetle case study

Planning forest management relies on predicting insect outbreaks such as mountain pine beetle, particularly in the intermediate‐term future, e.g., 5‐year. Machine‐learning algorithms are potential solutions to this challenging problem due to their many successes across a variety of prediction tasks. However, there are many subtle challenges in applying them: identifying the best learning models and the best subset of available covariates (including time lags) and properly evaluating the models to avoid misleading performance‐measures. We systematically address these issues in predicting the chance of a mountain pine beetle outbreak in the Cypress Hills area and seek models with the best performance at predicting future 1‐, 3‐, 5‐ and 7‐year infestations. We train nine machine‐learning models, including two generalized boosted regression trees (GBM) that predict future 1‐ and 3‐year infestations with 92% and 88% AUC, and two novel mixed models that predict future 5‐ and 7‐year infestations with 86% and 84% AUC, respectively. We also consider forming the train and test datasets by splitting the original dataset randomly rather than using the appropriate year‐based approach and show that this may obtain models that score high on the test dataset but low in practice, resulting in inaccurate performance evaluations. For example, a k‐nearest neighbor model with the actual performance of 68% AUC, scores the misleadingly high 78% on a test dataset obtained from a random split, but the more accurate 66% on a year‐based split. We then investigate how the prediction accuracy varies with respect to the provided history length of the covariates and find that neural network and naive Bayes, predict more accurately as history‐length increases, particularly for future 1‐ and 3‐year predictions, and roughly the same holds with GBM. Our approach is applicable to other invasive species. The resulting predictors can be used in planning forest and pest management and planning sampling locations in field studies.     

SARS‐CoV‐2 structural coverage map reveals viral protein assembly,mimicry, and hijacking mechanisms

We modeled 3D structures of all SARS‐CoV‐2 proteins, generating 2,060 models that span 69% of the viral proteome and provide details not available elsewhere. We found that ˜6% of the proteome mimicked human proteins, while ˜7% was implicated in hijacking mechanisms that reverse post‐translational modifications, block host translation, and disable host defenses; a further ˜29% self‐assembled into heteromeric states that provided insight into how the viral replication and translation complex forms. To make these 3D models more accessible, we devised a structural coverage map, a novel visualization method to show what is—and is not—known about the 3D structure of the viral proteome. We integrated the coverage map into an accompanying online resource (https://aquaria.ws/covid) that can be used to find and explore models corresponding to the 79 structural states identified in this work. The resulting Aquaria‐COVID resource helps scientists use emerging structural data to understand the mechanisms underlying coronavirus infection and draws attention to the 31% of the viral proteome that remains structurally unknown or dark.     

hu.MAP 2.0: integration of over 15,000 proteomic experiments builds a global compendium of human multiprotein assemblies

A general principle of biology is the self‐assembly of proteins into functional complexes. Characterizing their composition is, therefore, required for our understanding of cellular functions. Unfortunately, we lack knowledge of the comprehensive set of identities of protein complexes in human cells. To address this gap, we developed a machine learning framework to identify protein complexes in over 15,000 mass spectrometry experiments which resulted in the identification of nearly 7,000 physical assemblies. We show our resource, hu.MAP 2.0, is more accurate and comprehensive than previous state of the art high‐throughput protein complex resources and gives rise to many new hypotheses, including for 274 completely uncharacterized proteins. Further, we identify 253 promiscuous proteins that participate in multiple complexes pointing to possible moonlighting roles. We have made hu.MAP 2.0 easily searchable in a web interface (http://humap2.proteincomplexes.org/), which will be a valuable resource for researchers across a broad range of interests including systems biology, structural biology, and molecular explanations of disease.     

Identifying modules of cooperating cancer drivers

Identifying cooperating modules of driver alterations can provide insights into cancer etiology and advance the development of effective personalized treatments. We present Cancer Rule Set Optimization (CRSO) for inferring the combinations of alterations that cooperate to drive tumor formation in individual patients. Application to 19 TCGA cancer types revealed a mean of 11 core driver combinations per cancer, comprising 2–6 alterations per combination and accounting for a mean of 70% of samples per cancer type. CRSO is distinct from methods based on statistical co‐occurrence, which we demonstrate is a suboptimal criterion for investigating driver cooperation. CRSO identified well‐studied driver combinations that were not detected by other approaches and nominated novel combinations that correlate with clinical outcomes in multiple cancer types. Novel synergies were identified in NRAS‐mutant melanomas that may be therapeutically relevant. Core driver combinations involving NFE2L2 mutations were identified in four cancer types, supporting the therapeutic potential of NRF2 pathway inhibition. CRSO is available at https://github.com/mikekleinsgit/CRSO/.     

DOMINO: a network‐based active module identification algorithm with reduced rate of false calls

Algorithms for active module identification (AMI) are central to analysis of omics data. Such algorithms receive a gene network and nodes'' activity scores as input and report subnetworks that show significant over‐representation of accrued activity signal (“active modules”), thus representing biological processes that presumably play key roles in the analyzed conditions. Here, we systematically evaluated six popular AMI methods on gene expression and GWAS data. We observed that GO terms enriched in modules detected on the real data were often also enriched on modules found on randomly permuted data. This indicated that AMI methods frequently report modules that are not specific to the biological context measured by the analyzed omics dataset. To tackle this bias, we designed a permutation‐based method that empirically evaluates GO terms reported by AMI methods. We used the method to fashion five novel AMI performance criteria. Last, we developed DOMINO, a novel AMI algorithm, that outperformed the other six algorithms in extensive testing on GE and GWAS data. Software is available at https://github.com/Shamir‐Lab.     

Human pluripotent stem cell registry: Operations,role and current directions

The human plutiripotent stem cell registry (hPSCreg) is a global database for human embryonic and induced pluripotent stem cells (hESC, hiPSC). The publicly accessible Registry (https://hpscreg.eu) was set up to provide a transparent resource of quality‐assessed hPSC lines as well as to increase reproducibility of research and interoperability of data.ObjectivesIn this review, we describe the establishment of the Registry and its mission, its development into a knowledgebase for hPSC and the current status of hPSC‐focussed databases. The data categories available in hPSCreg are detailed. In addition, sharing and hurdles to data sharing on a global level are described.ConclusionsAn outlook is provided on the establishment of digital representatives of donors using hybrids of data and hPSC‐based biological models, and how this can also be used to reposition databases as mediators between donors and researchers.

hPSCreg as a data hub for pluripotent stem cells: Key utility and function.     

When we increase diversity in academia,we all win

Increasing diversity in academia is not just a matter of fairness but also improves science. It is up to individual scientists and research organisations to support underrepresented minorities. Subject Categories: S&S: Economics & Business, S&S: Ethics

There has been a large body of research on diversity in the workplace—in both academic and non‐academic settings—that highlights the benefits of an inclusive workplace. This is perhaps most clearly visible in industry where the rewards are immediate: A study by McKinsey showed that companies with a more diverse workforce perform better financially and by substantial margins, compared to their respective national industry medians (https://www.mckinsey.com/business-functions/organization/our-insights/why-diversity-matters#).It is easy to measure success in financial terms, but since there is no similar binary metric for research performance (https://sfdora.org), it is harder to quantify the rewards of workplace diversity in academic research. However, research shows that diversity actually provides research groups with a competitive edge in other quantifiable terms, such as citation counts (Powell, 2018), and the scientific process obviously benefits from diversity in perspectives. Bringing together individuals with different ways of thinking will allow us to solve more challenging scientific problems and lead to better decision‐making and leadership. Conversely, there is a direct cost to bias in recruitment, tenure, and promotion processes. When such processes are affected by bias—whether explicit or implicit—the whole organization is losing by not tapping into the wider range of skills and assets that could otherwise have been brought to the workplace. Promoting diversity in academia is therefore not simply an issue of equity, which in itself is a sufficient reason, but also a very practical question: how do we create a better work environment for our organization, both in terms of collegiality and in terms of performance?Notwithstanding the fact that there is now substantial awareness of the importance of diversity and that significant work is being invested into addressing the issue, the statistics do not look good. Despite a substantial improvement at the undergraduate and graduate student levels in the EU, women remain significantly underrepresented in research at the more senior levels (Directorate‐General for Research and Innovation European Commission, 2019). In addition, the lion’s share of diversity efforts, at least in Sweden where I work, is frequently focused on gender. Gender is clearly important, but other diversity axes with problematic biases deserve the same attention. As one example, while statistics on ethnic diversity is readily available for US Universities (Davis & Fry, 2019), this information is much harder to find in Europe. While there is an increased awareness of diversity at the student level, this does not necessarily translate into initiatives to support faculty diversity (Aragon & Hoskins, 2017). There are examples of progress and concrete actions on these fronts, including the Athena Swan Charter (https://www.ecu.ac.uk/equality-charters/athena-swan/), the more recent Race Equality Charter (https://www.advance-he.ac.uk/charters/race-equality-charter), and the EMBO journals that regularly analyze their decisions for gender bias. However, progress remains frustratingly slow. In 2019, the World Economic Forum suggested that, at the current rate of progress, the global gender gap will take 108 years to close (https://www.weforum.org/reports/the-global-gender-gap-report-2018). I worry that it may take even longer for other diversity axes since these receive far less attention.It is clear that there is a problem, but what can we do to address it? Perhaps one of the single most important contributions we can make as faculty is to address the implicit (subconscious) biases we all carry. Implicit bias will manifest itself in many ways: gender, ethnicity, socioeconomic status, or disability, just to mention a few. These are the easily identifiable ones, but implicit bias also extends to, for example, professional titles (seniority level), institutional affiliation and even nationality. These partialities affect our decision‐making—for example, in recruitment, tenure, promotion, and evaluation committees—and how we interact with each other.The “Matilda effect” (Rossiter, 1993), which refers to the diminishment of the value of contributions made by female researchers, is now well recognized, and it is not unique to gender (Ross, 2014). When we diminish the contributions of our colleagues, it affects how we evaluate them in competitive scenarios, and whether we put them forward for grants, prizes, recruitment, tenure, and so on. In the hypercompetitive environment that is academia today, even small and subtle injuries can tremendously amplify their negative impact on success, given the current reward system that appears to favor “fighters” over “collaborators”. Consciously working to correct for this, stepping back to rethink our first assessment, is imperative.Women and other minorities also frequently suffer from imposter syndrome, which can impact self‐confidence and make members of these groups less likely to self‐promote in the pursuit of prestigious funding, awards, and competitive career opportunities. This effect is further amplified by a globally mobile academic workforce who, when moving to new cultural contexts (whether locally or internationally), can be unaware of the unwritten rules that guide a department’s work environment and decision‐making processes. Here, mentoring can play a tremendous role in reducing barriers to success; however, for such mentorship to be productive, mentors need to be aware of the specific challenges faced by minorities in academia, as well as their own implicit biases (Hinton et al, 2020).Other areas where we, as individual academics, can contribute to a more diverse work environment include meeting cultures and decision‐making. Making sure that the members of decision‐making bodies have diverse composition so that a variety of views are represented is an important first step. One complication to bear in mind though is that implicit biases are not limited to individuals outside the group: A new UN report shows that almost 90% of people—both men and women—carry biases against women, which in turn is what contributes to the glass‐ceiling effect (United Nations Development Program, 2020). However, equally important is inclusiveness in the meeting culture. Studies from the business world show that even high‐powered women often struggle to speak up and be heard at meetings, and the onus for solving this is often passed back onto themselves. The same holds true for other minority groups, and in an academic setting, it extends to seminars and conferences. The next time you plan a meeting, think about the setting and layout. Who gets to talk? Why? Is the distribution of time given to participants representative of the composition of the meeting participants? If not, why not?As a final example of personal action, we can take: language matters (Ås, 1978). Even without malicious intent, there can be a big gap between what we say and mean, and how it comes across to the recipient. Some examples of this are given by Harrison and Tanner (Harrison & Tanner, 2018), who discuss microagressions in an academic setting and the underlying message one might be unintentionally sending. Microaggressions, when built up over a long period of time, and coming from different people, can significantly impact someone’s confidence and sense of self‐worth. Taking a step back and thinking about why we choose the language, we do is a vital part of creating an inclusive work environment.Addressing diversity challenges in academia is a highly complex multi‐faceted topic that is impossible to do justice in a short opinion piece. This is, therefore, just a small set of examples: By paying attention to our own biases and thinking carefully about how we interact with those around us, both in terms of the language we use and the working environments we create, we can personally contribute to improving both recruitment and retention of a diverse academic workforce. In addition, it is crucial to break the culture of silence and to speak up when we see others committing micro‐ or not so microaggressions or otherwise contributing to a hostile environment. There is a substantial amount of work that needs to be done, at both the individual and organization levels, before we have a truly inclusive academic environment. However, this is not a reason to not do it, and if each of us contributes, we can accelerate this change to a better and more equitable future, while all winning from the benefits of diversity.     

Ascochyta rabiei: A threat to global chickpea production

The necrotrophic fungus Ascochyta rabiei causes Ascochyta blight (AB) disease in chickpea. A. rabiei infects all aerial parts of the plant, which results in severe yield loss. At present, AB disease occurs in most chickpea‐growing countries. Globally increased incidences of A. rabiei infection and the emergence of new aggressive isolates directed the interest of researchers toward understanding the evolution of pathogenic determinants in this fungus. In this review, we summarize the molecular and genetic studies of the pathogen along with approaches that are helping in combating the disease. Possible areas of future research are also suggested.Taxonomykingdom Mycota, phylum Ascomycota, class Dothideomycetes, subclass Coelomycetes, order Pleosporales, family Didymellaceae, genus Ascochyta, species rabiei. Primary host A. rabiei survives primarily on Cicer species.Disease symptoms A. rabiei infects aboveground parts of the plant including leaves, petioles, stems, pods, and seeds. The disease symptoms first appear as watersoaked lesions on the leaves and stems, which turn brown or dark brown. Early symptoms include small circular necrotic lesions visible on the leaves and oval brown lesions on the stem. At later stages of infection, the lesions may girdle the stem and the region above the girdle falls off. The disease severity increases at the reproductive stage and rounded lesions with concentric rings, due to asexual structures called pycnidia, appear on leaves, stems, and pods. The infected pod becomes blighted and often results in shrivelled and infected seeds.Disease management strategiesCrop failures may be avoided by judicious practices of integrated disease management based on the use of resistant or tolerant cultivars and growing chickpea in areas where conditions are least favourable for AB disease development. Use of healthy seeds free of A. rabiei, seed treatments with fungicides, and proper destruction of diseased stubbles can also reduce the fungal inoculum load. Crop rotation with nonhost crops is critical for controlling the disease. Planting moderately resistant cultivars and prudent application of fungicides is also a way to combat AB disease. However, the scarcity of AB‐resistant accessions and the continuous evolution of the pathogen challenges the disease management process.Useful websites https://www.ndsu.edu/pubweb/pulse‐info/resourcespdf/Ascochyta%20blight%20of%20chickpea.pdf https://saskpulse.com/files/newsletters/180531_ascochyta_in_chickpeas‐compressed.pdf http://www.pulseaus.com.au/growing‐pulses/bmp/chickpea/ascochyta‐blight http://agriculture.vic.gov.au/agriculture/pests‐diseases‐and‐weeds/plant‐diseases/grains‐pulses‐and‐cereals/ascochyta‐blight‐of‐chickpea http://www.croppro.com.au/crop_disease_manual/ch05s02.php https://www.northernpulse.com/uploads/resources/722/handout‐chickpeaascochyta‐nov13‐2011.pdf http://oar.icrisat.org/184/1/24_2010_IB_no_82_Host_Plant https://www.crop.bayer.com.au/find‐crop‐solutions/by‐pest/diseases/ascochyta‐blight     

Nadeshiko revisited: The situation of women in Japanese research and the measures taken to increase their representation

The Japanese government has enacted measures to increase the representation of women in research; the situation is improving but there is still much to do. Subject Categories: S&S: Careers & Training, S&S: History & Philosophy of Science, S&S: Ethics

Japanese parents are understandably proud that their 15‐year‐old boys and girls do equally well in the Programme for International Student Assessment (PISA). In 2018, Japanese girls ranked second and third in Science and Mathematics, respectively, among the 40 participating countries, and Japanese boys ranked first in both subjects (https://data.oecd.org/japan.htm). However, Japanese boys and girls face different expectations and take different career paths as they grow up. In this commentary, we discuss how this affects the situation of female scientists in Japan. We start with the proportion of women in academic research and describe the problems they currently face. We underscore the tremendous measures developed and administered by the Japanese government to increase the participation and proportion of women in research. Finally, we mention an emerging grassroots initiative that is currently being implemented. We suggest that female empowerment may be one of the most promising strategies to improve the situation of women in the Japanese scientific community.     

A study of inhibitors of d-glycero-β-d-manno-heptose-1-phosphate adenylyltransferase from Burkholderia pseudomallei as a potential antibiotic target

d-Glycero-β-d-manno-heptose-1-phosphate adenylyltransferase from Burkholderia pseudomallei (BpHldC) is the fourth enzyme in the ADP‐l‐glycero‐β‐d‐manno‐heptose biosynthesis pathway producing a lipopolysaccharide core. Therefore, BpHldC is an anti-melioidosis target. Three ChemBridge compounds purchased from ChemBridge Corporation (San Diego, CA) were found to have an effective inhibitory activity on BpHldC. Interestingly, ChemBridge 7929959 was the most effective compound due to the presence of the terminal benzyl group. The enzyme kinetic study revealed that most of them show mixed type inhibitory modes against ATP and βG1P. The induced-fit docking indicated that the medium affinity of ChemBridge 7929959 is originated from its benzyl group occupying the substrate-binding pocket of BpHldC. The inhibitory role of terminal aromatic groups was proven with ChemBridge 7570508. Combined with the previous study, ChemBridge 7929959 is found to work as a dual inhibitor against both HldC and HddC. Therefore, three ChemBridge compounds can be developed as a potent anti-melioidosis agent with a novel inhibitory concept.     

PhenoSpace: A Shiny application to visualize trait data in the phenotypic space of the global spectrum of plant form and function

1. A recent analysis of variation in six major traits conducted on a large worldwide sample of vascular plant species showed that three‐quarters of trait variation was captured by a two‐dimensional global spectrum of plant form and function (“global spectrum” hereafter). We developed the PhenoSpace application, whose aim is to visualize and export the position of any individual/population/species in the phenotypic space of the global spectrum.
2. PhenoSpace is a Shiny application that helps users to manipulate and visualize data pertaining to the global spectrum of plant form and function. It is freely accessible at the following URL: https://shiny.cefe.cnrs.fr/PhenoSpace/.
3. PhenoSpace has three main functionalities. First, it allows users to visualize the phenotypic space of the global spectrum using different combinations of traits and growth forms. Second, trait data from any new user‐defined dataset can be projected onto the phenotypic space of the global spectrum, provided that at least two of the six traits are available. Finally, figures produced and loadings of the imported data on the PCA axes can be downloaded, allowing users to conduct further analyses.
4. PhenoSpace fulfills the practical goal of positioning plants in the phenotypic space of the global spectrum, making it possible to compare trait variation at any level of organization against the worldwide background. This serves a major aim of comparative plant ecology, which is to put specific sets of individuals, populations or species into a broader context, facilitating comparison and synthesis of results across different continents and environments using relevant indicators of plant design and function.