Behavioral fingerprints predict insecticide and anthelmintic mode of action

Systematic approaches to machine learning models for predicting pesticide toxicity

Pesticides play an important role in modern agriculture by protecting crops from pests and diseases. However, the negative consequences of pesticides, such as environmental contamination and adverse effects on human and ecological health, underscore the importance of accurate toxicity predictions. To address this issue, artificial intelligence models have emerged as valuable methods for predicting the toxicity of organic compounds. In this review article, we explore the application of machine learning (ML) for pesticide toxicity prediction. This review provides a detailed summary of recent developments, prediction models, and datasets used for pesticide toxicity prediction. In this analysis, we compared the results of several algorithms that predict the harmfulness of various classes of pesticides. Furthermore, this review article identified emerging trends and areas for future direction, showcasing the transformative potential of machine learning in promoting safer pesticide usage and sustainable agriculture.

Environmental carcinogen benzo[a]pyrene alters neutral lipid storage via a cyp-35A2 mediated pathway in Caenorhabditis elegans

Polycyclic aromatic hydrocarbons (PAHs), in particular benzo [a]pyrene (BaP), have been identified as carcinogenic components of tobacco smoke. In mammals, the toxicological response to BaP-diol-epoxide is driven by cytochrome P450 (CYP1A1), a pathway which is absent in Caenorhabditis elegans. In contrast, in worms prominently the CYP-35 enzyme family seems to be induced after BaP exposure. In C. elegans, BaP exposure reduces the accumulation of lysosomal neutral lipids in a dose dependent manner and the deletion of cyp-35A2 results in a significant elevation of neutral lipid metabolism. A cyp-35A2:mCherry;unc-47:GFP dual-labelled reporter strain facilitated the identification of three potential upstream regulators that drive BaP metabolism in worms, namely elt-2, nhr-49 and fos-1. This newly described reporter line is a powerful resource for future large-scale RNAi regarding toxicology and lipid metabolism screens.

A reductionist paradigm for high-throughput behavioural fingerprinting in Drosophila melanogaster

Understanding how the brain encodes behaviour is the ultimate goal of neuroscience and the ability to objectively and reproducibly describe and quantify behaviour is a necessary milestone on this path. Recent technological progresses in machine learning and computational power have boosted the development and adoption of systems leveraging on high-resolution video recording to track an animal pose and describe behaviour in all four dimensions. However, the high temporal and spatial resolution that these systems offer must come as a compromise with their throughput and accessibility. Here, we describe coccinella, an open-source reductionist framework combining high-throughput analysis of behaviour using real-time tracking on a distributed mesh of microcomputers (ethoscopes) with resource-lean statistical learning (HCTSA/Catch22). Coccinella is a reductionist system, yet outperforms state-of-the-art alternatives when exploring the pharmacobehaviour in Drosophila melanogaster.

Worm Generator: A System for High-Throughput in Vivo Screening

Large-scale screening of molecules in organisms requires high-throughput and cost-effective evaluating tools during preclinical development. Here, a novel in vivo screening strategy combining hierarchically structured biohybrid triboelectric nanogenerators (HB-TENGs) arrays with computational bioinformatics analysis for high-throughput pharmacological evaluation using Caenorhabditis elegans is described. Unlike the traditional methods for behavioral monitoring of the animals, which are laborious and costly, HB-TENGs with micropillars are designed to efficiently convert animals' behaviors into friction deformation and result in a contact-separation motion between two triboelectric layers to generate electrical outputs. The triboelectric signals are recorded and extracted to various bioinformation for each screened compound. Moreover, the information-rich electrical readouts are successfully demonstrated to be sufficient to predict a drug's identity by multiple-Gaussian-kernels-based machine learning methods. This proposed strategy can be readily applied to various fields and is especially useful in in vivo explorations to accelerate the identification of novel therapeutics.

wrmXpress: A modular package for high-throughput image analysis of parasitic and free-living worms

Advances in high-throughput and high-content imaging technologies require concomitant development of analytical software capable of handling large datasets and generating relevant phenotypic measurements. Several tools have been developed to analyze drug response phenotypes in parasitic and free-living worms, but these are siloed and often limited to specific instrumentation, worm species, and single phenotypes. No unified tool exists to analyze diverse high-content phenotypic imaging data of worms and provide a platform for future extensibility. We have developed wrmXpress, a unified framework for analyzing a variety of phenotypes matched to high-content experimental assays of free-living and parasitic nematodes and flatworms. We demonstrate its utility for analyzing a suite of phenotypes, including motility, development/size, fecundity, and feeding, and establish the package as a platform upon which to build future custom phenotypic modules. We show that wrmXpress can serve as an analytical workhorse for anthelmintic screening efforts across schistosomes, filarial nematodes, and free-living model nematodes and holds promise for enabling collaboration among investigators with diverse interests.

Making sense of sensory behaviors in vector-borne helminths

Migrations performed by helminths are impressive and diverse, and accumulating evidence shows that many are controlled by sophisticated sensory programs. The migrations of vector-borne helminths are particularly complex, requiring precise, stage-specific regulation. We review the contrasting states of knowledge on snail-borne schistosomes and mosquito-borne filarial nematodes. Rich observational data exist for the chemosensory behaviors of schistosomes, while the molecular sensory pathways in nematodes are well described. Recent investigations on the molecular mechanisms of sensation in schistosomes and filarial nematodes have revealed some features conserved within their respective phyla, but adaptations correlated with parasitism are pronounced. Technological developments are likely to extend these advances, and we forecast how these technologies may be applied.

Comprehensive analysis of locomotion dynamics in the protochordate Ciona intestinalis reveals how neuromodulators flexibly shape its behavioral repertoire

Vertebrate nervous systems can generate a remarkable diversity of behaviors. However, our understanding of how behaviors may have evolved in the chordate lineage is limited by the lack of neuroethological studies leveraging our closest invertebrate relatives. Here, we combine high-throughput video acquisition with pharmacological perturbations of bioamine signaling to systematically reveal the global structure of the motor behavioral repertoire in the Ciona intestinalis larvae. Most of Ciona’s postural variance can be captured by 6 basic shapes, which we term “eigencionas.” Motif analysis of postural time series revealed numerous stereotyped behavioral maneuvers including “startle-like” and “beat-and-glide.” Employing computational modeling of swimming dynamics and spatiotemporal embedding of postural features revealed that behavioral differences are generated at the levels of motor modules and the transitions between, which may in part be modulated by bioamines. Finally, we show that flexible motor module usage gives rise to diverse behaviors in response to different light stimuli.

Vertebrate nervous systems can generate a remarkable diversity of behaviors, but how did these evolve in the chordate lineage? A study of the protochordate Ciona intestinalis reveals novel insights into how a simple chordate brain uses neuromodulators to control its behavioral repertoire.

Image recognition based on deep learning in Haemonchus contortus motility assays

Poor efficacy of some anthelmintics and rising concerns about the widespread drug resistance have highlighted the need for new drug discovery. The parasitic nematode Haemonchus contortus is an important model organism widely used for studies of drug resistance and drug screening with the current gold standard being the motility assay. We applied a deep learning approach Mask R-CNN for analysing motility videos containing varying rates of motile worms and compared it to other commonly used algorithms with different levels of complexity, namely the Wiggle Index and the Wide Field-of-View Nematode Tracking Platform. Mask R-CNN consistently outperformed the other algorithms in terms of the detection of worms as well as the precision of motility forecasts, having a mean absolute percentage error of 7.6% and a mean absolute error of 5.6% for the detection and motility forecasts, respectively. Using Mask R-CNN for motility assays confirmed the common problem with algorithms that use non-maximum suppression in detecting overlapping objects, which negatively impacts the overall precision. The use of intersect over union as a measure of the classification of motile / non-motile instances had an overall accuracy of 89%, indicating that it is a viable alternative to previously used methods based on movement characteristics, such as body bends. In comparison to the existing methods evaluated here, Mask R-CNN performed better and we anticipate that this method will broaden the number of possible approaches to video analysis of worm motility.

Megapixel camera arrays enable high-resolution animal tracking in multiwell plates

Tracking small laboratory animals such as flies, fish, and worms is used for phenotyping in neuroscience, genetics, disease modelling, and drug discovery. An imaging system with sufficient throughput and spatiotemporal resolution would be capable of imaging a large number of animals, estimating their pose, and quantifying detailed behavioural differences at a scale where hundreds of treatments could be tested simultaneously. Here we report an array of six 12-megapixel cameras that record all the wells of a 96-well plate with sufficient resolution to estimate the pose of C. elegans worms and to extract high-dimensional phenotypic fingerprints. We use the system to study behavioural variability across wild isolates, the sensitisation of worms to repeated blue light stimulation, the phenotypes of worm disease models, and worms' behavioural responses to drug treatment. Because the system is compatible with standard multiwell plates, it makes computational ethological approaches accessible in existing high-throughput pipelines.

Physics-informed deep learning characterizes morphodynamics of Asian soybean rust disease

Medicines and agricultural biocides are often discovered using large phenotypic screens across hundreds of compounds, where visible effects of whole organisms are compared to gauge efficacy and possible modes of action. However, such analysis is often limited to human-defined and static features. Here, we introduce a novel framework that can characterize shape changes (morphodynamics) for cell-drug interactions directly from images, and use it to interpret perturbed development of Phakopsora pachyrhizi, the Asian soybean rust crop pathogen. We describe population development over a 2D space of shapes (morphospace) using two models with condition-dependent parameters: a top-down Fokker-Planck model of diffusive development over Waddington-type landscapes, and a bottom-up model of tip growth. We discover a variety of landscapes, describing phenotype transitions during growth, and identify possible perturbations in the tip growth machinery that cause this variation. This demonstrates a widely-applicable integration of unsupervised learning and biophysical modeling.

High-content approaches to anthelmintic drug screening

Most anthelmintics were discovered through in vivo screens using animal models of infection. Developing in vitro assays for parasitic worms presents several challenges. The lack of in vitro life cycle culture protocols requires harvesting worms from vertebrate hosts or vectors, limiting assay throughput. Once worms are removed from the host environment, established anthelmintics often show no obvious phenotype - raising concerns about the predictive value of many in vitro assays. However, with recent progress in understanding how anthelmintics subvert host-parasite interactions, and breakthroughs in high-content imaging and machine learning, in vitro assays have the potential to discern subtle cryptic parasite phenotypes. These may prove better endpoints than conventional in vitro viability assays.