Search for: All records

The availability of large datasets of organism images combined with advances in artificial intelligence (AI) has significantly enhanced the study of organisms through images, unveiling biodiversity patterns and macro-evolutionary trends. However, existing machine learning (ML)-ready organism datasets have several limitations. First, these datasets often focus on species classification only, overlooking tasks involving visual traits of organisms. Second, they lack detailed visual trait annotations, like pixel-level segmentation, that are crucial for in-depth biological studies. Third, these datasets predominantly feature organisms in their natural habitats, posing challenges for aquatic species like fish, where underwater images often suffer from poor visual clarity, obscuring critical biological traits. This gap hampers the study of aquatic biodiversity patterns which is necessary for the assessment of climate change impacts, and evolutionary research on aquatic species morphology. To address this, we introduce the Fish-Visual Trait Analysis (Fish-Vista) dataset—a large, annotated collection of about 80K fish images spanning 3000 different species, supporting several challenging and biologically relevant tasks including species classification, trait identification, and trait segmentation. These images have been curated through a sophisticated data processing pipeline applied to a cumulative set of images obtained from various museum collections. Fish-Vista ensures that visual traits of images are clearly visible, and provides fine-grained labels of various visual traits present in each image. It also offers pixel-level annotations of 9 different traits for about 7000 fish images, facilitating additional trait segmentation and localization tasks. The ultimate goal of Fish-Vista is to provide a clean, carefully curated, high-resolution dataset that can serve as a foundation for accelerating biological discoveries using advances in AI. Finally, we provide a comprehensive analysis of state-of-the-art deep learning techniques on Fish-Vista. 

The misfolding, aggregation, and spread of tau protein fibrils underlie tauopathies, a diverse class of neurodegenerative diseases for which effective treatments remain elusive. Among these are corticobasal dementia (CBD) and progressive supranuclear palsy (PSP), canonical examples of 4-repeat (4R) tauopathies characterized by tau isoforms exclusively with four microtubule-binding repeat domains. We target this 4R tau isoform-specific mechanism by focusing on misfolded tau’s distinctive stem-loop-stem structural motif formed by the junction of the 4R-defining alternatively spliced exon and the adjacent constitutive exon. A synthetic peptide based on this stem-loop-stem sequence can induce aggregation and spread in an isoform-specific manner. Here, we develop a protein-like polymer (PLP) in which multiple copies of this synthetic peptide form a brush-like structure capable of preventing tau aggregation by binding and capping fibril endsin vitro, in human brain organoids, and in cellular models with an EC50 of 105 ± 14 nM. PLPs demonstrate robust activity against fibrils derived from CBD and PSP patient brains and a PS19 mouse tauopathy model. Previous tau-targeted treatments have primarily focused on broad tau clearance, aggregation inhibition, or microtubule stabilization, often lacking isoform specificity and precision. In contrast, this approach targets the 4R tau isoform’s unique structural motif, offering a tailored therapeutic intervention for diseases like CBD and PSP. Supported by prior studies showing blood-brain barrier penetrance and safety profiles, this tau-binding PLP offers a promising translational path toward clinical applications in tauopathy treatment. 

In this paper, we extend the dataset statistics, model benchmarks, and performance analysis for the recently published KABR dataset, an in situ dataset for ungulate behavior recognition using aerial footage from the Mpala Research Centre in Kenya. The dataset comprises video footage of reticulated giraffes (lat. Giraffa reticulata), Plains zebras (lat. Equus quagga), and Grévy’s zebras (lat. Equus grevyi) captured using a DJI Mavic 2S drone. It includes both spatiotemporal (i.e., mini-scenes) and behavior annotations provided by an expert behavioral ecologist. In total, KABR has more than 10 hours of annotated video. We extend the previous work in four key areas by: (i) providing comprehensive dataset statistics to reveal new insights into the data distribution across behavior classes and species; (ii) extending the set of existing benchmark models to include a new state-of-the-art transformer; (iii) investigating weight initialization strategies and exploring whether pretraining on human action recognition datasets is transferable to in situ animal behavior recognition directly (i.e., zero-shot) or as initialization for end-to-end model training; and (iv) performing a detailed statistical analysis of the performance of these models across species, behavior, and formally defined segments of the long-tailed distribution. The KABR dataset addresses the limitations of previous datasets sourced from controlled environments, offering a more authentic representation of natural animal behaviors. This work marks a significant advancement in the automatic analysis of wildlife behavior, leveraging drone technology to overcome traditional observational challenges and enabling a more nuanced understanding of animal interactions in their natural habitats. The dataset is available at https://kabrdata.xyz 

Peptide-brush polymers generated by graft-through living polymerization of peptide-modified monomers exhibit high proteolytic stability, therapeutic efficacy, and potential as functional tandem repeat protein mimetics. Prior work has focused on polymers generated from structurally disordered peptides that lack defined conformations. To obtain insight into how the structure of these polymers is influenced by the folding of their peptide sidechains, a set of polymers with varying degrees of polymerization was prepared from peptide monomers that adopt α-helical secondary structure for comparison to those having random coil structures. Circular dichroism and nuclear magnetic resonance spectroscopy confirm the maintenance of the secondary structure of the constituent peptide when polymerized. Small-angle X-ray scattering (SAXS) studies reveal the solution-phase conformation of PLPs in different solvent environments. In particular, X-ray scattering shows that modulation of solvent hydrophobicity, as well as hydrogen bonding patterns of the peptide sidechain, plays an important role in the degree of globularity and conformation of the overall polymer, with polymers of helical peptide brushes showing less spherical compaction in conditions where greater helicity is observed. These structural insights into peptide brush folding and polymer conformation inform the design of these proteomimetic materials with promise for controlling and predicting their artificial fold and morphology 

Abstract Positional information encoded in signaling molecules is essential for early patterning in the prosensory domain of the developing cochlea. The sensory epithelium, the organ of Corti, contains an exquisite repeating pattern of hair cells and supporting cells. This requires precision in the morphogen signals that set the initial radial compartment boundaries, but this has not been investigated. To measure gradient formation and morphogenetic precision in developing cochlea, we developed a quantitative image analysis procedure measuring SOX2 and pSMAD1/5/9 profiles in mouse embryos at embryonic day (E)12.5, E13.5, and E14.5. Intriguingly, we found that the pSMAD1/5/9 profile forms a linear gradient up to the medial ~ 75% of the PSD from the pSMAD1/5/9 peak in the lateral edge during E12.5 and E13.5. This is a surprising activity readout for a diffusive BMP4 ligand secreted from a tightly constrained lateral region since morphogens typically form exponential or power-law gradient shapes. This is meaningful for gradient interpretation because while linear profiles offer the theoretically highest information content and distributed precision for patterning, a linear morphogen gradient has not yet been observed. Furthermore, this is unique to the cochlear epithelium as the pSMAD1/5/9 gradient is exponential in the surrounding mesenchyme. In addition to the information-optimized linear profile, we found that while pSMAD1/5/9 is stable during this timeframe, an accompanying gradient of SOX2 shifts dynamically. Last, through joint decoding maps of pSMAD1/5/9 and SOX2, we see that there is a high-fidelity mapping between signaling activity and position in the regions that will become Kölliker’s organ and the organ of Corti. Mapping is ambiguous in the prosensory domain precursory to the outer sulcus. Altogether, this research provides new insights into the precision of early morphogenetic patterning cues in the radial cochlea prosensory domain. 

Background The decision making process undertaken during wildfire responses is complex and prone to uncertainty. In the US, decisions federal land managers make are influenced by numerous and often competing factors. Aims To assess and validate the presence of decision factors relevant to the wildfire decision making context that were previously known and to identify those that have emerged since the US federal wildfire policy was updated in 2009. Methods Interviews were conducted across the US while wildfires were actively burning to elucidate time-of-fire decision factors. Data were coded and thematically analysed. Key results Most previously known decision factors as well as numerous emergent factors were identified. Conclusions To contextualise decision factors within the decision making process, we offer a Wildfire Decision Framework that has value for policy makers seeking to improve decision making, managers improving their process and wildfire social science researchers. Implications Managers may gain a better understanding of their decision environment and use our framework as a tool to validate their deliberations. Researchers may use these data to help explain the various pressures and influences modern land and wildfire managers experience. Policy makers and agencies may take institutional steps to align the actions of their staff with desired wildfire outcomes.