Automatic strawberry leaf scorch severity estimation via faster R-CNN and few-shot learning

Using deep learning to estimate strawberry leaf scorch severity often achieves unsatisfactory results when a strawberry leaf image contains complex background information or multi-class diseased leaves and the number of annotated strawberry leaf images is limited. To solve these issues, in this paper, we propose a two-stage method including object detection and few-shot learning to estimate strawberry leaf scorch severity. In the first stage, Faster R-CNN is used to mark the location of strawberry leaf patches, where each single strawberry leaf patch is clipped from original strawberry leaf images to compose a new strawberry leaf patch dataset. In the second stage, the Siamese network trained on the new strawberry leaf patch dataset is used to identify the strawberry leaf patches and then estimate the severity of the original strawberry leaf scorch images according to the multi-instance learning concept. Experimental results from the first stage show that Faster R-CNN achieves better mAP in strawberry leaf patch detection than other object detection networks, at 94.56%. Results from the second stage reveal that the Siamese network achieves an accuracy of 96.67% in the identification of strawberry disease leaf patches, which is higher than the Prototype network. Comprehensive experimental results indicate that compared with other state-of-the-art models, our proposed two-stage method comprising the Faster R-CNN (VGG16) and Siamese networks achieves the highest estimation accuracy of 96.67%. Moreover, our trained two-stage model achieves an estimation accuracy of 88.83% on a new dataset containing 60 strawberry leaf images taken in the field, which indicates its excellent generalization ability.     

BEEtag: A Low-Cost,Image-Based Tracking System for the Study of Animal Behavior and Locomotion

A fundamental challenge common to studies of animal movement, behavior, and ecology is the collection of high-quality datasets on spatial positions of animals as they change through space and time. Recent innovations in tracking technology have allowed researchers to collect large and highly accurate datasets on animal spatiotemporal position while vastly decreasing the time and cost of collecting such data. One technique that is of particular relevance to the study of behavioral ecology involves tracking visual tags that can be uniquely identified in separate images or movie frames. These tags can be located within images that are visually complex, making them particularly well suited for longitudinal studies of animal behavior and movement in naturalistic environments. While several software packages have been developed that use computer vision to identify visual tags, these software packages are either (a) not optimized for identification of single tags, which is generally of the most interest for biologists, or (b) suffer from licensing issues, and therefore their use in the study of animal behavior has been limited. Here, we present BEEtag, an open-source, image-based tracking system in Matlab that allows for unique identification of individual animals or anatomical markers. The primary advantages of this system are that it (a) independently identifies animals or marked points in each frame of a video, limiting error propagation, (b) performs well in images with complex backgrounds, and (c) is low-cost. To validate the use of this tracking system in animal behavior, we mark and track individual bumblebees (Bombus impatiens) and recover individual patterns of space use and activity within the nest. Finally, we discuss the advantages and limitations of this software package and its application to the study of animal movement, behavior, and ecology.     

Automatic detection of fish and tracking of movement for ecology

1. Animal movement studies are conducted to monitor ecosystem health, understand ecological dynamics, and address management and conservation questions. In marine environments, traditional sampling and monitoring methods to measure animal movement are invasive, labor intensive, costly, and limited in the number of individuals that can be feasibly tracked. Automated detection and tracking of small‐scale movements of many animals through cameras are possible but are largely untested in field conditions, hampering applications to ecological questions.
2. Here, we aimed to test the ability of an automated object detection and object tracking pipeline to track small‐scale movement of many individuals in videos. We applied the pipeline to track fish movement in the field and characterize movement behavior. We automated the detection of a common fisheries species (yellowfin bream, Acanthopagrus australis) along a known movement passageway from underwater videos. We then tracked fish movement with three types of tracking algorithms (MOSSE, Seq‐NMS, and SiamMask) and evaluated their accuracy at characterizing movement.
3. We successfully detected yellowfin bream in a multispecies assemblage (F1 score =91%). At least 120 of the 169 individual bream present in videos were correctly identified and tracked. The accuracies among the three tracking architectures varied, with MOSSE and SiamMask achieving an accuracy of 78% and Seq‐NMS 84%.
4. By employing this integrated object detection and tracking pipeline, we demonstrated a noninvasive and reliable approach to studying fish behavior by tracking their movement under field conditions. These cost‐effective technologies provide a means for future studies to scale‐up the analysis of movement across many visual monitoring systems.

     

Do small-scale exclosure/enclosure experiments predict the effects of large-scale extirpation of freshwater migratory fauna?

A variety of theoretical and empirical studies indicate that the abilities of small-scale experiments to predict responses to large-scale perturbations vary. Small-scale experiments often do not predict the directions of large-scale responses, and relatively few empirical studies have examined whether small-scale experiments predict the magnitudes of large-scale responses. Here we present an empirical example of small-scale manipulations predicting not only the directions but also the magnitudes of the effects of whole-catchment, decades-long decimation of migratory freshwater shrimp populations. In streams of Puerto Rico (USA), we used arena sizes of < 2 m² in 1- to 4-week exclosure/enclosure experiments. Effects of small-scale experiments largely matched those of large-scale shrimp loss above dams for a variety of response variables (abiotic and biotic factors including epilithic fine sediments, algae and organic matter, and invertebrate grazers, detritivores, and predators). The results of our extrapolation contrast with studies of small- versus large-scale perturbations in the temperate zone. Our findings are likely explained by: a set of response variables that are more dominated by within-patch processes than exchange processes, an experimental manipulation that encompassed the characteristic scales of response variables, our use of open arenas lacking cage artifacts, and/or our combination of two distinct experimental approaches (exclosures and enclosures). Based on our study design, we suggest that extrapolation across experimental scales can be greatly enhanced by embedding open arenas within large-scale conditions that represent all treatment levels. Electronic Supplementary Material Supplementary material is available to authorised users in the online version of this article at .     

Individual differences in the perception of probability

In recent studies of humans estimating non-stationary probabilities, estimates appear to be unbiased on average, across the full range of probability values to be estimated. This finding is surprising given that experiments measuring probability estimation in other contexts have often identified conservatism: individuals tend to overestimate low probability events and underestimate high probability events. In other contexts, repulsive biases have also been documented, with individuals producing judgments that tend toward extreme values instead. Using extensive data from a probability estimation task that produces unbiased performance on average, we find substantial biases at the individual level; we document the coexistence of both conservative and repulsive biases in the same experimental context. Individual biases persist despite extensive experience with the task, and are also correlated with other behavioral differences, such as individual variation in response speed and adjustment rates. We conclude that the rich computational demands of our task give rise to a variety of behavioral patterns, and that the apparent unbiasedness of the pooled data is an artifact of the aggregation of heterogeneous biases.     

Linear Superposition and Prediction of Bacterial Promoter Activity Dynamics in Complex Conditions

Bacteria often face complex environments. We asked how gene expression in complex conditions relates to expression in simpler conditions. To address this, we obtained accurate promoter activity dynamical measurements on 94 genes in E. coli in environments made up of all possible combinations of four nutrients and stresses. We find that the dynamics across conditions is well described by two principal component curves specific to each promoter. As a result, the promoter activity dynamics in a combination of conditions is a weighted average of the dynamics in each condition alone. The weights tend to sum up to approximately one. This weighted-average property, called linear superposition, allows predicting the promoter activity dynamics in a combination of conditions based on measurements of pairs of conditions. If these findings apply more generally, they can vastly reduce the number of experiments needed to understand how E. coli responds to the combinatorially huge space of possible environments.     

Syndromes or Flexibility: Behavior during a Life History Transition of a Coral Reef Fish

The theory of behavioral syndromes focuses on quantifying variation in behavior within and among individual organisms and attempts to account for the maintenance of differences in behavior that occur in a consistent manner among individuals. Behavioral syndromes have potentially important ecological consequences (e.g. survivorship tradeoffs) and can be shaped by population dynamics through selective mortality. Here, we search for any evidence for consistency of behavior across situations in juveniles of a common damselfish, Pomacentrus amboinensis (Pomacentridae) at the transition between larval habitats in the plankton and juvenile habitats on the reef. Naïve fish leaving the pelagic phase to settle on reefs were caught by light traps and their behaviors observed using similar methods across three different situations (small aquaria, large aquaria, field setting); all of which represent low risk and well-sheltered environments. Seven behavioral traits were compared within and among individuals across situations to determine if consistent behavioral syndromes existed. No consistency was found in any single or combination of behavioral traits for individuals across all situations. We suggest that high behavioral flexibility is likely beneficial for newly-settled fish at this ontogenetic transition and it is possible that consistent behavioral syndromes are unlikely to emerge in juveniles until environmental experience is gained or certain combinations of behaviors are favored by selective mortality.     

Sweet potato leaf detection in a natural scene based on faster R-CNN with a visual attention mechanism and DIoU-NMS

Accurate detection of plant leaves is a meaningful and challenging task for developing smart agricultural systems. To improve the performance of detecting plant leaves in natural scenes containing severe occlusion, overlapping, or shape variation, we developed an in situ sweet potato leaf detection method based on a modified Faster R-CNN framework and visual attention mechanism. First, a convolutional block attention module was added to the backbone network to enhance and extract critical features of leaf images by fusing cross-channel information and spatial information. Subsequently, the DIoU-NMS algorithm was adopted to modify the regional proposal network by replacing the original NMS. DIoU-NMS was utilized to reduce missed and incorrect detection in scenes of densely distributed leaves by considering the targets' overlap ratio, distance, and scale. The proposed leaf detection method was tested and evaluated on sweet potato plant images collected in agricultural fields. In the datasets, sweet potato leaves were presented in various sizes and poses, and a large proportion of leaves were occluded or overlapped with each other. The experimental results showed that the proposed leaf detection method outperforms state-of-the-art object detection methods. The mean average precision of the proposed method reached 95.7%, which was 2.9% higher than that of the original Faster R-CNN and 7.0% higher than that of YOLOv5. The proposed method achieved promising performance in detecting dense leaves or occluded leaves and could provide key techniques for applications in smart agriculture and ecological monitoring, such as growth monitoring or plant phenotyping.     

Beyond the response—High throughput behavioral analyses to link genome to phenome in Caenorhabditis elegans

The development and application of methods for automated behavioral analysis have revolutionized behavioral genetics across model organisms. In this review we summarize the history of automated behavioral analysis in the nematode Caenorhabditis elegans. We highlight recent studies of learning and memory to exemplify just how complex the genetic and neural circuit mechanisms underlying a seemingly simple single behavioral response can be. We finish by looking forward at the exciting prospects of combing genomic technologies with connectomic and phenomic level measurements.     

An Efficient Multiple-Testing Adjustment for eQTL Studies that Accounts for Linkage Disequilibrium between Variants

Methods for multiple-testing correction in local expression quantitative trait locus (cis-eQTL) studies are a trade-off between statistical power and computational efficiency. Bonferroni correction, though computationally trivial, is overly conservative and fails to account for linkage disequilibrium between variants. Permutation-based methods are more powerful, though computationally far more intensive. We present an alternative correction method called eigenMT, which runs over 500 times faster than permutations and has adjusted p values that closely approximate empirical ones. To achieve this speed while also maintaining the accuracy of permutation-based methods, we estimate the effective number of independent variants tested for association with a particular gene, termed M_eff, by using the eigenvalue decomposition of the genotype correlation matrix. We employ a regularized estimator of the correlation matrix to ensure M_eff is robust and yields adjusted p values that closely approximate p values from permutations. Finally, using a common genotype matrix, we show that eigenMT can be applied with even greater efficiency to studies across tissues or conditions. Our method provides a simpler, more efficient approach to multiple-testing correction than existing methods and fits within existing pipelines for eQTL discovery.     

Biological modeling of complex chemotaxis behaviors for C. elegans under speed regulation—a dynamic neural networks approach

In this paper, the modeling of several complex chemotaxis behaviors of C. elegans is explored, which include food attraction, toxin avoidance, and locomotion speed regulation. We first model the chemotaxis behaviors of food attraction and toxin avoidance separately. Then, an integrated chemotaxis behavioral model is proposed, which performs the two chemotaxis behaviors simultaneously. The novelty and the uniqueness of the proposed chemotaxis behavioral models are characterized by several attributes. First, all the chemotaxis behavioral model sare on biological basis, namely, the proposed chemotaxis behavior models are constructed by extracting the neural wire diagram from sensory neurons to motor neurons, where sensory neurons are specific for chemotaxis behaviors. Second, the chemotaxis behavioral models are able to perform turning and speed regulation. Third, chemotaxis behaviors are characterized by a set of switching logic functions that decide the orientation and speed. All models are implemented using dynamic neural networks (DNN) and trained using the real time recurrent learning (RTRL) algorithm. By incorporating a speed regulation mechanism, C. elegans can stop spontaneously when approaching food source or leaving away from toxin. The testing results and the comparison with experiment results verify that the proposed chemotaxis behavioral models can well mimic the chemotaxis behaviors of C. elegans in different environments.     

WormPose: Image synthesis and convolutional networks for pose estimation in C. elegans

An important model system for understanding genes, neurons and behavior, the nematode worm C. elegans naturally moves through a variety of complex postures, for which estimation from video data is challenging. We introduce an open-source Python package, WormPose, for 2D pose estimation in C. elegans, including self-occluded, coiled shapes. We leverage advances in machine vision afforded from convolutional neural networks and introduce a synthetic yet realistic generative model for images of worm posture, thus avoiding the need for human-labeled training. WormPose is effective and adaptable for imaging conditions across worm tracking efforts. We quantify pose estimation using synthetic data as well as N2 and mutant worms in on-food conditions. We further demonstrate WormPose by analyzing long (∼ 8 hour), fast-sampled (∼ 30 Hz) recordings of on-food N2 worms to provide a posture-scale analysis of roaming/dwelling behaviors.     

High-Resolution Video Tracking of Locomotion in Adult Drosophila Melanogaster

Flies provide an important model for studying complex behavior due to the plethora of genetic tools available to researchers in this field. Studying locomotor behavior in Drosophila melanogaster relies on the ability to be able to quantify changes in motion during or in response to a given task. For this reason, a high-resolution video tracking system, such as the one we describe in this paper, is a valuable tool for measuring locomotion in real-time. Our protocol involves the use of an initial air pulse to break the flies momentum, followed by a thirty second filming period in a square chamber. A tracking program is then used to calculate the instantaneous speed of each fly within the chamber in 10 msec increments. Analysis software then compiles this data, and outputs a variety of parameters such as average speed, max speed, time spent in motion, acceleration, etc. This protocol will discuss proper feeding and management of flies for behavioral tasks, handling flies without anesthetization or immobilization, setting up a controlled environment, and running the assay from start to finish.Open in a separate windowClick here to view.^{(55M, flv)}     

Warming, plant phenology and the spatial dimension of trophic mismatch for large herbivores

Temporal advancement of resource availability by warming in seasonal environments can reduce reproductive success of vertebrates if their own reproductive phenology does not also advance with warming. Indirect evidence from large-scale analyses suggests, however, that migratory vertebrates might compensate for this by tracking phenological variation across landscapes. Results from our two-year warming experiment combined with seven years of observations of plant phenology and offspring production by caribou (Rangifer tarandus) in Greenland, however, contradict evidence from large-scale analyses. At spatial scales relevant to the foraging horizon of individual herbivores, spatial variability in plant phenology was reduced--not increased--by both experimental and observed warming. Concurrently, offspring production by female caribou declined with reductions in spatial variability in plant phenology. By highlighting the spatial dimension of trophic mismatch, these results reveal heretofore unexpected adverse consequences of climatic warming for herbivore population ecology.     

Statistical modeling for selecting housekeeper genes

There is a need for statistical methods to identify genes that have minimal variation in expression across a variety of experimental conditions. These 'housekeeper' genes are widely employed as controls for quantification of test genes using gel analysis and real-time RT-PCR. Using real-time quantitative RT-PCR, we analyzed 80 primary breast tumors for variation in expression of six putative housekeeper genes (MRPL19 (mitochondrial ribosomal protein L19), PSMC4 (proteasome (prosome, macropain) 26S subunit, ATPase, 4), SF3A1 (splicing factor 3a, subunit 1, 120 kDa), PUM1 (pumilio homolog 1 (Drosophila)), ACTB (actin, beta) and GAPD (glyceraldehyde-3-phosphate dehydrogenase)). We present appropriate models for selecting the best housekeepers to normalize quantitative data within a given tissue type (for example, breast cancer) and across different types of tissue samples.     

Contributions and promise of human behavioral genetics

Human behavioral genetics has contributed greatly to our understanding of human behavioral development. Twin, family, and adoption studies have shown that genetic effects are ubiquitous and that both genes and environments contribute to individual differences in behavior. The unique ability of behavioral genetic methods to separate genetic from environmental effects has also led to important discoveries about how the environment works in development and to the elucidation of the complex ways environments and genes interact across the life span. Although quantitative methods have been the mainstay of the field of human behavioral genetics since Galton's time, the Human Genome Project and advances in molecular genetics are providing new tools and promise as we enter the 21st century. Thus the future of human behavioral genetics lies in the cross-disciplinary exchanges and collaborations that will increasingly occur in the years to come among quantitative and molecular scientists who work with both animal and human systems. This research may someday culminate in an understanding of the biological basis of behavior that spans from how the brain develops and functions to a grasp of how genes influence thought at the molecular level.     

Large scale maximum average power multiple inference on time-course count data with application to RNA-seq analysis

Experiments that longitudinally collect RNA sequencing (RNA-seq) data can provide transformative insights in biology research by revealing the dynamic patterns of genes. Such experiments create a great demand for new analytic approaches to identify differentially expressed (DE) genes based on large-scale time-course count data. Existing methods, however, are suboptimal with respect to power and may lack theoretical justification. Furthermore, most existing tests are designed to distinguish among conditions based on overall differential patterns across time, though in practice, a variety of composite hypotheses are of more scientific interest. Finally, some current methods may fail to control the false discovery rate. In this paper, we propose a new model and testing procedure to address the above issues simultaneously. Specifically, conditional on a latent Gaussian mixture with evolving means, we model the data by negative binomial distributions. Motivated by Storey (2007) and Hwang and Liu (2010), we introduce a general testing framework based on the proposed model and show that the proposed test enjoys the optimality property of maximum average power. The test allows not only identification of traditional DE genes but also testing of a variety of composite hypotheses of biological interest. We establish the identifiability of the proposed model, implement the proposed method via efficient algorithms, and demonstrate its good performance via simulation studies. The procedure reveals interesting biological insights, when applied to data from an experiment that examines the effect of varying light environments on the fundamental physiology of the marine diatom Phaeodactylum tricornutum.     

Flux module decomposition for parameter estimation in a multiple-feedback loop model of biochemical networks

Computer simulation is an important technique to capture the dynamics of biochemical networks. Since few quantitative values are measured in vivo, the values for unmeasured parameters should be estimated so that the simulation agrees with the experimental data. Considering the sparsity and error rates of experimentally measured data, the first thing is not to find a numerically exact and global solution but to explore a variety of the plausible parameter solutions. To find many plausible parameter solutions without any biases, we developed the two-phase search (TPS) method. However, calculation complexity makes it hard for TPS to optimize a large-scale dynamic model. In this study divide-and-conquer methods are used to solve this problem. The flux module decomposition (FMD) is first proposed that separates a complex, large-scale dynamic model into multiple flux modules without deteriorating its basic control architectures. FMD is combined with TPS, named FMD-TPS, to find many plausible parameter solutions for a dynamic model. To demonstrate the feasibility of FMD-TPS, it is applied to the E. coli ammonia assimilation system that consists of multiple-feedback loops. The variability of the solutions is verified by measuring the space distribution of the parameter solution vectors and by defining the binary vectors checking the consistency with biological behaviors. Compared with non-decomposition methods, FMD-TPS efficiently explored a variety of plausible parameter solutions that reproduce the dynamic behaviors in vivo.     

Are behavioral responses to eyespots in sticklebacks influenced by the visual environment? An experimental examination

Eyespots are taxonomically widespread color patterns consisting of large concentric rings that are commonly assumed to protect prey by influencing the behaviors of predators. Although there is ample experimental evidence supporting an anti‐predator function of eyespots in terrestrial animals, whether eyespots have a similar deterring function in aquatic animals remains unclear. Furthermore, studies in terrestrial systems suggest that the protective function of eyespots depends on ambient light conditions where predators encounter them, but this effect has never been tested in aquatic environments. Here, we examine how eyespots influence behavioral responses in an aquatic environment under different visual environments, using laboratory‐reared three‐spined sticklebacks (Gasterosteus aculeatus) as model predators. Specifically, we experimentally examined behavioral responses of sticklebacks toward artificial prey patterns (control vs. eyespots) under two different light environment treatments (low vs. high). We found that eyespots did not postpone attacks from sticklebacks. However, sticklebacks approaching eyespots stopped more frequently than sticklebacks approaching prey items with a control pattern. Sticklebacks were (marginally) slower to attack prey in the low‐light treatment, but the light level did not influence stickleback behavioral responses toward eyespots. We conclude that eyespots can modulate some behaviors of an aquatic predator, albeit with a different functional role from that previously demonstrated in terrestrial species.