Working Title – Subject to Revision
This is a live document, and will be subject to change as the project progresses
This project is intended to support training and continuing education in appraising goat mammary systems. The ADGA linear appraisal system converts visual evaluation into standardized numerical measurements that are associated with ease of milking, udder durability, mastitis resistance, and long-term productivity. Because these scores are used in breeding and herd management decisions, consistent interpretation can affect breeding value estimation and long-term herd improvement.
This project is being developed in collaboration with the American Dairy Goat Association (ADGA) and UC Davis.
The ADGA is the primary organization that sets standards and maintains records for dairy goats in the U.S. One of its key programs is the Linear Appraisal System, which evaluates goats based on physical traits. In this system, trained appraisers evaluate physical traits related to mammary function, and dairy characteristics.
These evaluations are used to support:
- breeding decisions
- herd management
- genetic evaluation and research
Each trait is scored on a 1 - 50 scale based on observable physical variation. In the ADGA system, these scores are treated as a “linear” description of a trait, meaning they describe a range of biological variation rather than assigning a ranking of good or bad.
“The term ‘linear’ in a linear appraisal system refers to the fact that traits are rated on a linear scale that goes from one biological extreme for that trait to the other.”
— ADGA Linear Appraisal Booklet
For example, a trait score may describe differences in udder height, teat placement, or leg structure. The goal of the system is to create a more standardized and consistent way to describe dairy goat conformation across herds and appraisers.
However, interpreting the relationship between numeric scores and physical form still depends heavily on visual experience and in-person training. Outside formal appraisal sessions, it can be difficult to consistently visualize what different trait scores represent in practice, which can contribute to variation between appraisers.
The goal is to build an application that takes linear appraisal scores for mammary traits and converts them into a visual rear-view model of the udder.
This project focuses specifically on the mammary system from the rear view (the udder and teat structures responsible for milk production), a component of dairy goat evaluation.
Key traits include:
- rear udder height
- udder depth
- udder arch
- medial support (udder split)
- teat placement and length
These traits are associated with milk production, udder durability, and susceptibility to injury or mastitis. For example, excessively deep udders may be more vulnerable to injury and infection due to their proximity to the hocks and ground.
Because these evaluations are used in breeding and herd management decisions, consistent interpretation of mammary traits is important for assessing long-term productivity and structural soundness in dairy goats. The scoring of these traits should be an approximate Gaussian distribution that represents the biological range seen across the goat population. In practice, appraisers estimate traits visually, which can lead to variation in scores. Specifically, a distribution that is too tight around the mean and therefore biologically unrepresentative.
This project explores whether linear appraisal traits can be translated into a visual rear-view udder model that could support appraisal training, score interpretation, and consistency between appraisers in and outside formal evaluation sessions.
The diagram below shows an example of how one trait (teat placement) is visualized in the ADGA Linear Appraisal system.
Source: American Dairy Goat Association (ADGA), Linear Appraisal materials (LABOOKLETALL_19.pdf).
This diagram is included as a reference to show how numeric scores correspond to physical traits. In the ADGA system, scores reflect differences in position, proportion, and structure (such as how centrally the teats are placed on the udder).
-- In focusing on the rear view of the udder, this project:
- converts numeric scores into a visual representation
- makes these traits easier to interpret
- helps check consistency in scoring
- provides a reference for training and validation
- support training outside of formal appraisal sessions
This project takes linear appraisal scores for mammary traits and converts them into a rear-view visual model of a dairy goat udder.
Users input trait values related to:
- udder height and depth
- udder shape and attachment
- medial support (udder split)
- teat placement and teat length
The system then uses those values to adjust different parts of the udder model and generate a corresponding visualization.
Different anatomical features are built as separate R scripts and combined through a Shiny application workflow. Rear leg and pelvic positioning are also included as reference points to help maintain proportional relationships within the model.
The current prototype does not yet directly translate official ADGA linear appraisal scores into finalized biological representations. Instead, the current system is being used to explore how trait-based parameters may be visually modeled and interpreted.
The image below shows the current visualization output generated from the Ag-GOAT mammary trait modeling workflow.
The current prototype allows users to adjust mammary trait parameters and generate a rear-view udder visualization that can be used as a visual reference during development and testing.
The diagram below outlines the current development workflow for the project, including the separation of mammary traits into modular R source functions, integration through app.R, and planned future development stages.
The designs below represent exploratory interface and visual design concepts for the Ag-GOAT application. They are intended to demonstrate possible layout structure, workflow organization, and visual styling choices for future development.
This project is developed in R and uses a Shiny-based interface.
Required packages:
install.packages(c("tidyverse", "shiny", "magick", "shinylive"))
[Google Drive] (https://drive.google.com/drive/folders/1k2zZalMFZtyAQk7a1ZUKVJFEten8hQ5k?usp=sharing)
data/
├── Goat Pictures/ Images with scale reference (ruler)
├── rear udder image library/ Rear-view udder images
├──2025 LA Data_Cleaned.xlsx Cleaned linear appraisal dataset with goat trait scores and standardized variables for analysis and modeling
└──2025 LA Data_Uncleaned.xlsx Raw linear appraisal dataset containing original recorded trait scores and animal information Images of miniature goats taken from rear
scoping_documentation/
├── Linear Modeling Tool Grant_DataLab_2026.docx Project proposal outlining goals, methods, and planned development
├── Scoping Meeting Notes Notes from initial project discussions and planning meetings
└── scoping_document.docx Formal project scope, roles, responsibilities, and deliverables
2025Linear-SOPDraft1.pdf Documentation of ADGA Linear Appraisal traits and scoring system
Data Inventory Metadata table describing datasets used in the project, including size, source, and structure
GOAT Readme.md Overview of the project and objectives
Linear 2025.pptx Data analysis and visualizations of 2025 appraisal scores, including trait distributions and inter-appraiser variability
Meeting Notes Notes from meetings with project lead and principal investigators
Student Meeting Notes Notes from internal student team meetings
R/ R source code
├── data_cleaning.R Data preprocessing
├── leg_curve.R Function: rear leg / hock reference (used for proportional scoring)
├── pelvic_curve.R Function: pelvic arch reference (anchor for udder traits)
├── udder_curve.R Function: udder shape (height, depth, arch)
├── medial_curve.R Function: medial suspensory ligament (udder support)
├── teats_curve.R Function: teat placement and length (rear view)
└── ui_teats.R Prototype Shiny UI for testing visualization
data/ Will contain rear udder reference images for input parameterization
docs/ Supporting documents
├── team_agreement.md Team workflow guidelines and collaboration expectations
└── README_TEATS.md
images/
├── Ag-GOAT Figma.png UI design prototype
├── goat_workflow.png Diagram showing how project scripts and application components connect
├── goat_timeline_workflow.png Visual timeline of project development stages and milestones
└── teat_placement.png ADGA diagram showing teat placement scoring scale
.gitignore Paths Git should ignore
README.md This file
- Goat Data Dictionary
Reference document describing datasets, variables, trait abbreviations, and source information used throughout the project.
Fauna Smith
Assistant Professor
flsmith@ucdavis.edu
Ben Rupchist
Manager, Goat Teaching & Research Facility
barupchis@ucdavis.edu
Nora Manring
Student Assistant, Goat Teaching & Research Facility
ezmanring@ucdavis.edu
Colton Baumler
PhD Candidate
ccbaumler@ucdavis.edu
Nick Ulle
Senior Statistician
naulle@ucdavis.edu
Camila Chicatto
Undergraduate Student
cchicatto@ucdavis.edu
Hilary Choi
Undergraduate Student
hfchoi@ucdavis.edu
Justin Huang
Undergraduate Student
jjhhwang@ucdavis.edu
Rashmit Shrestha
Undergraduate Student
rlshrestha@ucdavis.edu
Odelyn Xie
Undergraduate Student
delxie@ucdavis.edu