Context

The Stanford Dogs dataset contains images of 120 breeds of dogs from around the world. This dataset has been built using images and annotation from ImageNet for the task of fine-grained image categorization. It was originally collected for fine-grain image categorization, a challenging problem as certain dog breeds have near identical features or differ in colour and age. I have used only images, so this does not contain any labels .

  Content

Number of… See the full description on the dataset page: https://huggingface.co/datasets/ksaml/Stanford_dogs.

Over the last century, dogs have been increasingly used to detect rare and elusive species or traces of them. The use of wildlife detection dogs (WDD) is particularly well established in North America, Europe and Oceania, and projects deploying them have increased worldwide. However, if they are to make a significant contribution to conservation and management, their strengths, abilities, and limitations should be fully identified. We reviewed the use of WDD with particular focus on the breeds used in different countries and for various targets, as well as their overall performance compared to other methods, by developing and analysing a database of 1220 publications, including 916 scientific ones, covering 2464 individual cases - most of them (1840) scientific. With the worldwide increase in the use of WDD, associated tasks have changed and become much more diverse. Since 1930, reports exist for 62 countries and 407 animal, 42 plant, 26 fungi and 6 bacteria species. Altogether, 108 FCI...

The Stanford Dogs dataset contains 20,580 images of 120 classes of dogs from around the world, which are divided into 12,000 images for training and 8,580 images for testing.

BackgroundCanine rabies is a neglected disease causing 55,000 human deaths worldwide per year, and 99% of all cases are transmitted by dog bites. In N'Djaména, the capital of Chad, rabies is endemic with an incidence of 1.71/1,000 dogs (95% C.I. 1.45–1.98). The gold standard of rabies diagnosis is the direct immunofluorescent antibody (DFA) test, requiring a fluorescent microscope. The Centers for Disease Control and Prevention (CDC, Atlanta, United States of America) developed a histochemical test using low-cost light microscopy, the direct rapid immunohistochemical test (dRIT).Methodology/Principal FindingsWe evaluated the dRIT in the Chadian National Veterinary Laboratory in N'Djaména by testing 35 fresh samples parallel with both the DFA and dRIT. Additional retests (n = 68 in Chad, n = 74 at CDC) by DFA and dRIT of stored samples enhanced the power of the evaluation. All samples were from dogs, cats, and in one case from a bat. The dRIT performed very well compared to DFA. We found a 100% agreement of the dRIT and DFA in fresh samples (n = 35). Results of retesting at CDC and in Chad depended on the condition of samples. When the sample was in good condition (fresh brain tissue), we found simple Cohen's kappa coefficient related to the DFA diagnostic results in fresh tissue of 0.87 (95% C.I. 0.63–1) up to 1. For poor quality samples, the kappa values were between 0.13 (95% C.I. −0.15–0.40) and 0.48 (95% C.I. 0.14–0.82). For samples stored in glycerol, dRIT results were more likely to agree with DFA testing in fresh samples than the DFA retesting.Conclusion/SignificanceThe dRIT is as reliable a diagnostic method as the gold standard (DFA) for fresh samples. It has an advantage of requiring only light microscopy, which is 10 times less expensive than a fluorescence microscope. Reduced cost suggests high potential for making rabies diagnosis available in other cities and rural areas of Africa for large populations for which a capacity for diagnosis will contribute to rabies control.

IntroductionChronic kidney disease (CKD) in canines is a progressive condition characterized by a gradual decline in kidney function. There are significant gaps in understanding how CKD is managed in canines and the full extent of its impact. This study aimed to characterize disease management of CKD and its impact on dogs, their owners and the veterinary healthcare system in the United States of America (United States).MethodsData were drawn from the Adelphi Real World Canine CKD Disease Specific Programme™, a cross-sectional survey of veterinarians, pet owners and their dogs with CKD in the United States from December 2022 to January 2024. Veterinarians reported demographic, diagnostic, treatment, and healthcare utilization data, for dogs with CKD. Owners voluntarily completed questionnaires, providing data about their dog, as well as quality of life and work-related burden using the Dog Owners Quality of Life, and the Work Productivity and Activity Impairment questionnaires. Analyses were descriptive and Cohen’s Kappa was used to measure agreement between owners and veterinarians.ResultsA total of 117 veterinarians provided data for 308 dogs, of which 68 owners also reported information. Discrepancies in recognizing symptoms of CKD in dogs, particularly excessive water consumption and urination, were identified between veterinary professionals and owners. Interventions for managing CKD in dogs focused on controlling symptoms and supporting kidney function through dietary modifications and medication. Owners of dogs with CKD reported minimal impact to overall work and activity impairment (10 and 14%, respectively). At diagnosis, 78.6% of dogs were International Renal Interest Society Stage I-II, and 21.5% were Stage III-IV. Regardless of CKD stage, owners strongly agreed that ownership provided them with emotional support and companionship. Regarding veterinary healthcare utilization, 95% of dogs were seen in general veterinary practices.DiscussionThese findings emphasize the value of real-world evidence in enhancing our understanding of CKD in companion animals and informs future strategy for the real-world diagnosis and treatment of CKD. The results also provide insights to the potential burden experienced by owners of dogs with CKD.

Background

The first-ever, large-scale generative modeling research competition, Generative Dog Images, was held on Kaggle in the summer of 2019. Over 900+ teams participated and submitted a total of 10k+ generated samples, 1.6k of which were selected as the final submissions to rank on the private leaderboard. We are releasing the competition submissions as an effort to facilitate research on generative modeling metric design, particularly towards tackling the issue of detecting training sample memorization, intentional or not.

Content

Each competition submission consists of 10k generated samples of dog images from a generative model trained on the Stanford dogs dataset. As expected participants are incentivized to optimize for the objective and many exploited the insensitivity to training sample memorization issue of current popular generative modeling metrics (e.g. IS, FID). We provided manual labels of the type of intentional memorization technique adopted (if any) for each submission. Details regarding the labels can be found in the description of the labels.csv file. We also provided human-assessed image quality annotations for individual images.

Acknowledgements

Huge thanks to all the participants in the Generative Dog Images research competition for providing all the well-tuned models as well as feedback during the competition. The competition result analysis is published as a conference paper and if you find this dataset useful, please cite the following: @inproceedings{bai2021genmem, author = {Ching-Yuan Bai and Hsuan-Tien Lin and Colin Raffel and Wendy Chih-wen Kan}, title = {On Training Sample Memorization: Lessons from Benchmarking Generative Modeling with a Large-scale Competition}, booktitle = {Proceedings of the 27th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD)}, year = 2021, month = aug }

Inspiration

The Memorizaion-informed Fréchet Inception Distance (MiFID) was proposed and adopted as the benchmark metric during the competition to handle the training sample memorization issue. It works well in a competition setting but obvious flaws make it unideal in a general research setting.

Are there any other alternatives?

The large amount and great diversity of models in this dataset can serve as a testing ground for newly developed benchmark metrics.

Description:

This dataset contains comprehensive demographic, medical, pain-related, and personality data from 200 dogs recruited from a large referral small animal hospital in Switzerland. The dataset is part of the "Paws in Pain" project, designed to develop a multidimensional database aimed at advancing the assessment and understanding of canine pain. This database integrates various validated scales, including the Visual Analogue Scale (VAS), Glasgow Composite Measure Pain Scale (GCPS), and personality assessments using the Monash and PANAS questionnaires.

Dataset contents:

· Demographic data, including age, breed, sex, neuter status.

· Medical data, including primary diagnosis, medication administered, patient type (hospitalized, surgical, consultation, emergency).

· Pain-related data, including type of pain (such as acute, chronic, inflammatory, neuropathic, nociceptive), pain scores from VAS and GCPS.

· Personality data, including the item scores from Monash and Positive and Negative Activation Scale (PANAS) assessing personality traits like anxiety, fearfulness, and sociability.

· Additionally, video recordings have been collected from these dogs. Researchers interested in accessing these videos for collaboration projects are invited to contact the dataset authors.

Format: Excel (.xlsx)

Usage notes: This dataset can be utilized to assess and analyse questions related to canine pain such as correlations among canine pain levels, medical conditions, personality traits, breeds, diagnoses, and demographic variables. Researchers may use these data to explore relationships between pain states and personality, investigate breed-specific differences in pain expression, or examine the interaction between clinical diagnoses and pain severity.

Ethics and consent: Ethical approval for this study was granted by the University of Lincoln Ethics Committee under the approval reference UoL2024_18004. In compliance with the Swiss Animal Welfare Act (TSchG), this study did not require approval from the Committee for Animal Experimentation of the Canton of Bern (Switzerland), as the data were collected during routine diagnostic and therapeutic procedures. Dog owners provided consent for the use of their pets' data for research purposes.

Suggested citation: Bremhorst, A. (2025). Dog Pain Database: A Multidimensional Dataset for Investigating Canine Pain. Zenodo. DOI: 10.5281/zenodo.15303646.

License: Creative Commons Attribution 4.0 International (CC BY 4.0)

Funding and acknowledgments: This project was supported by an SNSF Spark Grant. Special thanks to the veterinary hospital and all dog owners who participated in this research.

Authors and contributors:

· Annika Bremhorst, University of Bern

· Claudia Spadavecchia, University of Bern

· Michelle Braghetti, University of Bern

· Anna Zamansky, University of Haifa

· Liat Vichman, University of Haifa

· Nareed Farhat, University of Haifa

· Daniel Mills, University of Lincoln

Introduction

This dataset supports Ye et al. 2024 Nature Communications. Please cite this dataset and paper if you use this resource. Please also see Ye et al. 2024 for the full DataSheet that accompanies this download, including the meta data for how to use this data is you want to compare model results on benchmark tasks. Below is just a summary. Also see the dataset licensing below.

Training Data

It consists of being trained together on the following datasets:

• AwA-Pose Quadruped dataset, see full details at (1).
• AnimalPose See full details at (2).
• AcinoSet See full details at (3).
• Horse-30 Horse-30 dataset, benchmark task is called Horse-10; See full details at (4).
• StanfordDogs See full details at (5, 6).
• AP-10K See full details at (7).
• iRodent We utilized the iNaturalist API functions for scraping observations with the taxon ID of Suborder Myomorpha (8). The functions allowed us to filter the large amount of observations down to the ones with photos under the CC BY-NC creative license. The most common types of rodents from the collected observations are Muskrat (Ondatra zibethicus), Brown Rat (Rattus norvegicus), House Mouse (Mus musculus), Black Rat (Rattus rattus), Hispid Cotton Rat (Sigmodon hispidus), Meadow Vole (Microtus pennsylvanicus), Bank Vole (Clethrionomys glareolus), Deer Mouse (Peromyscus maniculatus), White-footed Mouse (Peromyscus leucopus), Striped Field Mouse (Apodemus agrarius). We then generated segmentation masks over target animals in the data by processing the media through an algorithm we designed that uses a Mask Region Based Convolutional Neural Networks(Mask R-CNN) (9) model with a ResNet-50-FPN backbone (10), pretrained on the COCO datasets (11). The processed 443 images were then manually labeled with both pose annotations and segmentation masks. iRodent data is banked at https://zenodo.org/record/8250392.
• APT-36K See full details at (12).

https://images.squarespace-cdn.com/content/v1/57f6d51c9f74566f55ecf271/1690988780004-AG00N6OU1R21MZ0AU9RE/modelcard-SAQ.png?format=1500w" target="_blank" rel="noopener">Here is an image with a keypoint guide.

Ethical Considerations

• No experimental data was collected for this model; all datasets used are cited above.

Caveats and Recommendations

• Please note that each dataest was labeled by separate labs & separate individuals, therefore while we map names to a unified pose vocabulary, there will be annotator bias in keypoint placement (See Ye et al. 2024 for our Supplementary Note on annotator bias). You will also note the dataset is highly diverse across species, but collectively has more representation of domesticated animals like dogs, cats, horses, and cattle. We recommend if performance of a model trained on this data is not as good as you need it to be, first try video adaptation (see Ye et al. 2024), or fine-tune the weights with your own labeling.

License

Modified MIT.

Copyright 2023-present by Mackenzie Mathis, Shaokai Ye, and contributors.

Permission is hereby granted to you (hereafter "LICENSEE") a fully-paid, non-exclusive,
and non-transferable license for academic, non-commercial purposes only (hereafter “LICENSE”)
to use the "DATASET" subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial
portions of the Software:

This data or resulting software may not be used to harm any animal deliberately.

LICENSEE acknowledges that the DATASET is a research tool.
THE DATASET IS PROVIDED “AS IS”, WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING
BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE DATASET OR THE USE OR OTHER DEALINGS IN THE DATASET.

If this license is not appropriate for your application, please contact Prof. Mackenzie W. Mathis
(mackenzie@post.harvard.edu) for a commercial use license.

Please cite Ye et al if you use this DATASET in your work.

References

1. Prianka Banik, Lin Li, and Xishuang Dong. A novel dataset for keypoint detection of quadruped animals from images. ArXiv, abs/2108.13958, 2021
2. Jinkun Cao, Hongyang Tang, Haoshu Fang, Xiaoyong Shen, Cewu Lu, and Yu-Wing Tai. Cross-domain adaptation for animal pose estimation. 2019 IEEE/CVF International Conference on Computer Vision (ICCV), pages 9497–9506, 2019.
3. Daniel Joska, Liam Clark, Naoya Muramatsu, Ricardo Jericevich, Fred Nicolls, Alexander Mathis, Mackenzie W. Mathis, and Amir Patel. Acinoset: A 3d pose estimation dataset and baseline models for cheetahs in the wild. 2021 IEEE International Conference on Robotics and Automation (ICRA), pages 13901–13908, 2021.
4. Alexander Mathis, Thomas Biasi, Steffen Schneider, Mert Yuksekgonul, Byron Rogers, Matthias Bethge, and Mackenzie W Mathis. Pretraining boosts out-of-domain robustness for pose estimation. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision, pages 1859–1868, 2021.
5. Aditya Khosla, Nityananda Jayadevaprakash, Bangpeng Yao, and Li Fei-Fei. Novel dataset for fine-grained image categorization. In First Workshop on Fine-Grained Visual Categorization, IEEE Conference on Computer Vision and Pattern Recognition, Colorado Springs, CO, June 2011.
6. Benjamin Biggs, Thomas Roddick, Andrew Fitzgibbon, and Roberto Cipolla. Creatures great and smal: Recovering the shape and motion of animals from video. In Asian Conference on Computer Vision, pages 3–19. Springer, 2018.
7. Hang Yu, Yufei Xu, Jing Zhang, Wei Zhao, Ziyu Guan, and Dacheng Tao. Ap-10k: A benchmark for animal pose estimation in the wild. In Thirty-fifth Conference on Neural Information Processing Systems Datasets and Benchmarks Track (Round 2), 2021.
8. iNaturalist. OGBIF Occurrence Download. https://doi.org/10.15468/dl.p7nbxt. iNaturalist, July 2020
9. Kaiming He, Georgia Gkioxari, Piotr Dollár, and Ross Girshick. Mask r-cnn. In Proceedings of the IEEE international conference on computer vision, pages 2961–2969, 2017.
10. Tsung-Yi Lin, Piotr Dollár, Ross Girshick, Kaiming He, Bharath Hariharan, and Serge Belongie. Feature pyramid networks for object detection, 2016.
11. Tsung-Yi Lin, Michael Maire, Serge J. Belongie, Lubomir D. Bourdev, Ross B. Girshick, James Hays, Pietro Perona, Deva Ramanan, Piotr Doll’ar, and C. Lawrence Zitnick. Microsoft COCO: common objects in context. CoRR, abs/1405.0312, 2014
12. Yuxiang Yang, Junjie Yang, Yufei Xu, Jing Zhang, Long Lan, and Dacheng Tao. Apt-36k: A large-scale benchmark for animal pose estimation and tracking. Advances in Neural Information Processing Systems, 35:17301–17313, 2022

Versioning Note:

- V2 includes fixes to Stanford Dog data; it affected less than 1% of the data.

Survey Data. Upon enrollment in Darwin’s Ark (https://darwinsark.org), owners were asked to provide consent for participation and information about their dog’s approximate birth date, sex and spay/neuter status, suspected or known breed(s), purebred registration, and/or photograph. We presented owners with 22 surveys composed of 10-12 questions, for which any number or order can be answered. The majority offered response choices of agreement with statements presented or the frequency of behavior in question on a 5-point Likert scale. Among these surveys, 123 questions were sourced from published and validated canine behavioral and health surveys, including the Dog Personality Questionnaire (DPQ / DPQL), the Canine Health-related Quality of Life Survey (CHQLS), the Dog Impulsivity Assessment Scale (DIAS), the Canine Cognitive Dysfunction Rating scale (CCDR), the Certified Dog Trainer Test (CDTT, International Association of Canine Professionals), and the Dog Obesity Risk and Appetit...

The long-term monitoring of carnivore tracks in Doñana is part of a harmonised protocol for the Long-term Ecological Monitoring Program of Natural Resources and Processes targeting mammals' populations. The general aim of this protocol is to study the temporal evolution of the relative density of the main species of carnivores in the main habitats of the Doñana National Park. Tracks surveys were done annually after the first rains of the hydrological year, i.e. the first autumn rains, usually in October. Due to climate change, in recent years the rainy season has been delayed until the beginning of the year. This protocol has stablished in 2007 and it has done annually until the present (2022), except in 2021 when due to logistical problems no census was made. Censuses are carried out through 12 prefixed transects, with sand substrate, in Doñana National Park. Each transect consists of a 2 km of length and 1.5 m of width that is done by a car at a constant speed between 10 and 15 km/h. Transects are cleaned the day before of the census with a metal beam to facilitate the read of the tracks and to ensure that the foot prints were from the previous day. Each transect is repeated in three consecutive days, and during the transect the sand is cleaned for the next day. In the census an expert in mammals’ tracks identifies all the tracks, i.e. groups of carnivore foot prints, and he/she records them in Cybertracker. That way, tracks' information like coordinates, hour, species identification and observation was recorded; and also the information of each transect was recorded: surveyors, drivers, date, start and end (hour and coordinates). This method enables to calculate Kilometric Abundance Indexes (KAI) for each species and transect. In order to clarify all carnivore datasets, the data was separated by species, this allows concrete analysis by species. In this dataset domestic dog´s (Canis lupus familiaris) data is presented.

Description

The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS) contains 7356 files (total size: 24.8 GB). The dataset contains 24 professional actors (12 female, 12 male), vocalizing two lexically-matched statements in a neutral North American accent. Speech includes calm, happy, sad, angry, fearful, surprise, and disgust expressions, and song contains calm, happy, sad, angry, and fearful emotions. Each expression is produced at two levels of emotional intensity (normal, strong), with an additional neutral expression. All conditions are available in three modality formats: Audio-only (16bit, 48kHz .wav), Audio-Video (720p H.264, AAC 48kHz, .mp4), and Video-only (no sound). Note, there are no song files for Actor_18.

The RAVDESS was developed by Dr Steven R. Livingstone, who now leads the Affective Data Science Lab, and Dr Frank A. Russo who leads the SMART Lab.

Citing the RAVDESS

The RAVDESS is released under a Creative Commons Attribution license, so please cite the RAVDESS if it is used in your work in any form. Published academic papers should use the academic paper citation for our PLoS1 paper. Personal works, such as machine learning projects/blog posts, should provide a URL to this Zenodo page, though a reference to our PLoS1 paper would also be appreciated.

Academic paper citation

Livingstone SR, Russo FA (2018) The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS): A dynamic, multimodal set of facial and vocal expressions in North American English. PLoS ONE 13(5): e0196391. https://doi.org/10.1371/journal.pone.0196391.

Personal use citation

Include a link to this Zenodo page - https://zenodo.org/record/1188976

Commercial Licenses

Commercial licenses for the RAVDESS can be purchased. For more information, please visit our license page of fees, or contact us at ravdess@gmail.com.

Contact Information

If you would like further information about the RAVDESS, to purchase a commercial license, or if you experience any issues downloading files, please contact us at ravdess@gmail.com.

Example Videos

Watch a sample of the RAVDESS speech and song videos.

Emotion Classification Users

If you're interested in using machine learning to classify emotional expressions with the RAVDESS, please see our new RAVDESS Facial Landmark Tracking data set [Zenodo project page].

Construction and Validation

Full details on the construction and perceptual validation of the RAVDESS are described in our PLoS ONE paper - https://doi.org/10.1371/journal.pone.0196391.

The RAVDESS contains 7356 files. Each file was rated 10 times on emotional validity, intensity, and genuineness. Ratings were provided by 247 individuals who were characteristic of untrained adult research participants from North America. A further set of 72 participants provided test-retest data. High levels of emotional validity, interrater reliability, and test-retest intrarater reliability were reported. Validation data is open-access, and can be downloaded along with our paper from PLoS ONE.

Contents

Audio-only files

Audio-only files of all actors (01-24) are available as two separate zip files (~200 MB each):

• Speech file (Audio_Speech_Actors_01-24.zip, 215 MB) contains 1440 files: 60 trials per actor x 24 actors = 1440.
• Song file (Audio_Song_Actors_01-24.zip, 198 MB) contains 1012 files: 44 trials per actor x 23 actors = 1012.

Audio-Visual and Video-only files

Video files are provided as separate zip downloads for each actor (01-24, ~500 MB each), and are split into separate speech and song downloads:

• Speech files (Video_Speech_Actor_01.zip to Video_Speech_Actor_24.zip) collectively contains 2880 files: 60 trials per actor x 2 modalities (AV, VO) x 24 actors = 2880.
• Song files (Video_Song_Actor_01.zip to Video_Song_Actor_24.zip) collectively contains 2024 files: 44 trials per actor x 2 modalities (AV, VO) x 23 actors = 2024.

File Summary

In total, the RAVDESS collection includes 7356 files (2880+2024+1440+1012 files).

File naming convention

Each of the 7356 RAVDESS files has a unique filename. The filename consists of a 7-part numerical identifier (e.g., 02-01-06-01-02-01-12.mp4). These identifiers define the stimulus characteristics:

Filename identifiers

• Modality (01 = full-AV, 02 = video-only, 03 = audio-only).
• Vocal channel (01 = speech, 02 = song).
• Emotion (01 = neutral, 02 = calm, 03 = happy, 04 = sad, 05 = angry, 06 = fearful, 07 = disgust, 08 = surprised).
• Emotional intensity (01 = normal, 02 = strong). NOTE: There is no strong intensity for the 'neutral' emotion.
• Statement (01 = "Kids are talking by the door", 02 = "Dogs are sitting by the door").
• Repetition (01 = 1st repetition, 02 = 2nd repetition).
• Actor (01 to 24. Odd numbered actors are male, even numbered actors are female).

Filename example: 02-01-06-01-02-01-12.mp4

1. Video-only (02)
2. Speech (01)
3. Fearful (06)
4. Normal intensity (01)
5. Statement "dogs" (02)
6. 1st Repetition (01)
7. 12th Actor (12)
8. Female, as the actor ID number is even.

License information

The RAVDESS is released under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License, CC BY-NC-SA 4.0

Commercial licenses for the RAVDESS can also be purchased. For more information, please visit our license fee page, or contact us at ravdess@gmail.com.

Related Data sets

• RAVDESS Facial Landmark Tracking data set [Zenodo project page].

Primary open angle glaucoma (POAG) is a leading cause of blindness worldwide, with elevated intraocular pressure as an important risk factor. Increased resistance to outflow of aqueous humor through the trabecular meshwork causes elevated intraocular pressure, but the specific mechanisms are unknown. In this study, we used genome-wide SNP arrays to map the disease gene in a colony of Beagle dogs with inherited POAG to within a single 4 Mb locus on canine chromosome 20. The Beagle POAG locus is syntenic to a previously mapped human quantitative trait locus for intraocular pressure on human chromosome 19. Sequence capture and next-generation sequencing of the entire canine POAG locus revealed a total of 2,692 SNPs segregating with disease. Of the disease-segregating SNPs, 54 were within exons, 8 of which result in amino acid substitutions. The strongest candidate variant causes a glycine to arginine substitution in a highly conserved region of the metalloproteinase ADAMTS10. Western blotting revealed ADAMTS10 protein is preferentially expressed in the trabecular meshwork, supporting an effect of the variant specific to aqueous humor outflow. The Gly661Arg variant in ADAMTS10 found in the POAG Beagles suggests that altered processing of extracellular matrix and/or defects in microfibril structure or function may be involved in raising intraocular pressure, offering specific biochemical targets for future research and treatment strategies.