Digital and genetic tools to improve bones of laying hens

Sallam Mohammed Abdallah

doi:10.54612/a.q33hkb4fja

Digital and genetic tools to improve bones of laying hens

Bidragets översatta titel: Digitala och genetiska verktyg för att förbättra benhälsan hos värphöns

Sallam Mohammed Abdallah

Institutionen för husdjurens biovetenskaper

Publikation: Avhandling › Doktorsavhandling

25 Nedladdningar

Sammanfattning

Skelettbenskador är vanliga hos värphöns. Denna avhandling undersöker användningen av röntgen och genetiska som potentiella verktyg för att förbättra benhälsan. I en kohort av renrasiga höns undersöktes genomik för bensammansättning och styrka. Ett omvänt samband mellan benets innehåll av fetter och benstyrka identifierades, vilket visade på genetiska markörer kopplade till fettinnehållet som kan användas i aveln. Eftersom renrasiga höns används i aveln, studerades även benen hos hybridhöns, för att förbättra urvalet av renrasiga höns. Genetiken, inkl. SNP-effekter, för ben från hybridhöns i inredda burar och i system för frigående höns, befanns ha en noggrannhet på 0,42–0,65. I näst del röntgades hybridhöns upprepade gånger på en kommersiell gård. Mätningar av röntgentäthet i bröstbenskam och skenben, samt bröstbenets geometri, gjordes för att undersöka sambandet mellan bröstbensskador och bäckenstorlek vid obduktion. Hönor med en större bäckenhåla hade ökad risk för skador på bröstbenskammen. Ökat förhållande mellan bröstbenskammens längd och medeldjup vid 42, 55 och 68 veckors ålder var korrelerat med skador på bröstbenen. Höns med högre bentäthet i skenbenet vid 55 veckors ålder hade färre skador på bröstbenet. Med hjälp av datorseende och maskininlärning utvecklades algoritmer för att segmentera bröstbenskammen från röntgenbilder (med en noggrannhet på 0,88–0,90), och dessutom kunde bröstbenets geometri och bentäthet beräknas. Dessa mått visade på svag till måttlig ärftlighet, och på måttliga till starka genetiska korrelationer med bröstbensskador. Automatisering av röntgenanalys gör det möjligt att effektivt kunna fenotypa tusentals hönor på en begränsad tid. Avelsföretag kan använda de nya metoderna för att selektera höns med minskad risk för bröstbenskador.

Bidragets översatta titel	Digitala och genetiska verktyg för att förbättra benhälsan hos värphöns
Originalspråk	Engelska
Kvalifikation	Doktor i filosofi
Förlag	Swedish University of Agricultural Sciences
Tryckta ISBN	978-91-8046-447-5
Elektroniska ISBN	978-91-8046-497-0
DOI	https://doi.org/10.54612/a.q33hkb4fja
Status	Publicerad - 2025

Nyckelord

X-ray
bone
computer vision
deviations
fractures
genetics
genomics
keel
laying hens
machine learning
radiograph
tibia

SLU-serie

Acta Universitatis Agriculturae Sueciae

Åtkomst av dokument

10.54612/a.q33hkb4fja

FulltextSlutlig publicerad version, 6,02 MBLicens: CC BY

Andra filer och länkar

Permanent länk till denna sida

Citera det här

@phdthesis{e4e2492068ab4a74a57b2278926e4867,

title = "Digital and genetic tools to improve bones of laying hens",

abstract = "Bone damage is widespread in laying hens. This thesis explores the potential use of X-ray imaging and genetic tools to improve bone health. In a cohort of purebred hens, we investigated the genomics of bone composition and strength. An inverse relationship was found between bone lipids and strength, identifying genetic markers linked to lipid levels that could guide practical interventions to enhance bone strength. Since purebreds are typically selected to improve crossbred performance, we studied the bones of crossbreds and the potential use of crossbred data to inform purebred selection for various crossbred environments. The genetics, including SNP effects, of bones from crossbred hens in cage and cage-free housing were predicted with an accuracy of 0.42{\textendash}0.65, potentially aiding purebred selection. In another experiment, crossbred hens were repeatedly X-rayed on-farm, and both keel and tibia bones were measured on the X-ray images. X-ray measurements of keel and tibia density, as well as keel geometry, were tested for associations with post-mortem findings of keel damage and pelvic capacity. Larger pelvic capacity was associated with increased keel damage. A higher ratio of keel length to mid-depth at 42, 55, and 68 weeks of age was associated with greater deviations observed in dissected keels. A higher tibia radiographic density at 55 weeks of age was associated with fewer deviations and fractures. Automating X-ray analysis is crucial for efficiently phenotyping thousands of birds. Using deep learning and computer vision, we developed algorithms to segment the keel bone from whole-body X-ray images (with 0.88{\textendash}0.90 accuracy) and compute keel geometry and density metrics. These metrics showed weak to moderate heritability and moderate to strong genetic correlations with keel deviations and fractures, enabling quick phenotyping once birds are Xrayed. Breeding companies can use the new methods to select for reduced keel damage and optimize housing and nutrition strategies for better keel health.",

keywords = "X-ray, bone, computer vision, deviations, fractures, genetics, genomics, keel, laying hens, machine learning, radiograph, tibia, X-ray, bone, computer vision, deviations, fractures, genetics, genomics, keel, laying hens, machine learning, radiograph, tibia",

author = "\{Mohammed Abdallah\}, Sallam",

year = "2025",

doi = "10.54612/a.q33hkb4fja",

language = "English",

isbn = "978-91-8046-447-5",

series = "Acta Universitatis Agriculturae Sueciae",

publisher = "Swedish University of Agricultural Sciences",

number = "2025:12",

address = "Sweden",

}

TY - BOOK

T1 - Digital and genetic tools to improve bones of laying hens

AU - Mohammed Abdallah, Sallam

PY - 2025

Y1 - 2025

N2 - Bone damage is widespread in laying hens. This thesis explores the potential use of X-ray imaging and genetic tools to improve bone health. In a cohort of purebred hens, we investigated the genomics of bone composition and strength. An inverse relationship was found between bone lipids and strength, identifying genetic markers linked to lipid levels that could guide practical interventions to enhance bone strength. Since purebreds are typically selected to improve crossbred performance, we studied the bones of crossbreds and the potential use of crossbred data to inform purebred selection for various crossbred environments. The genetics, including SNP effects, of bones from crossbred hens in cage and cage-free housing were predicted with an accuracy of 0.42–0.65, potentially aiding purebred selection. In another experiment, crossbred hens were repeatedly X-rayed on-farm, and both keel and tibia bones were measured on the X-ray images. X-ray measurements of keel and tibia density, as well as keel geometry, were tested for associations with post-mortem findings of keel damage and pelvic capacity. Larger pelvic capacity was associated with increased keel damage. A higher ratio of keel length to mid-depth at 42, 55, and 68 weeks of age was associated with greater deviations observed in dissected keels. A higher tibia radiographic density at 55 weeks of age was associated with fewer deviations and fractures. Automating X-ray analysis is crucial for efficiently phenotyping thousands of birds. Using deep learning and computer vision, we developed algorithms to segment the keel bone from whole-body X-ray images (with 0.88–0.90 accuracy) and compute keel geometry and density metrics. These metrics showed weak to moderate heritability and moderate to strong genetic correlations with keel deviations and fractures, enabling quick phenotyping once birds are Xrayed. Breeding companies can use the new methods to select for reduced keel damage and optimize housing and nutrition strategies for better keel health.

AB - Bone damage is widespread in laying hens. This thesis explores the potential use of X-ray imaging and genetic tools to improve bone health. In a cohort of purebred hens, we investigated the genomics of bone composition and strength. An inverse relationship was found between bone lipids and strength, identifying genetic markers linked to lipid levels that could guide practical interventions to enhance bone strength. Since purebreds are typically selected to improve crossbred performance, we studied the bones of crossbreds and the potential use of crossbred data to inform purebred selection for various crossbred environments. The genetics, including SNP effects, of bones from crossbred hens in cage and cage-free housing were predicted with an accuracy of 0.42–0.65, potentially aiding purebred selection. In another experiment, crossbred hens were repeatedly X-rayed on-farm, and both keel and tibia bones were measured on the X-ray images. X-ray measurements of keel and tibia density, as well as keel geometry, were tested for associations with post-mortem findings of keel damage and pelvic capacity. Larger pelvic capacity was associated with increased keel damage. A higher ratio of keel length to mid-depth at 42, 55, and 68 weeks of age was associated with greater deviations observed in dissected keels. A higher tibia radiographic density at 55 weeks of age was associated with fewer deviations and fractures. Automating X-ray analysis is crucial for efficiently phenotyping thousands of birds. Using deep learning and computer vision, we developed algorithms to segment the keel bone from whole-body X-ray images (with 0.88–0.90 accuracy) and compute keel geometry and density metrics. These metrics showed weak to moderate heritability and moderate to strong genetic correlations with keel deviations and fractures, enabling quick phenotyping once birds are Xrayed. Breeding companies can use the new methods to select for reduced keel damage and optimize housing and nutrition strategies for better keel health.

KW - X-ray

KW - bone

KW - computer vision

KW - deviations

KW - fractures

KW - genetics

KW - genomics

KW - keel

KW - laying hens

KW - machine learning

KW - radiograph

KW - tibia

KW - X-ray

KW - bone

KW - computer vision

KW - deviations

KW - fractures

KW - genetics

KW - genomics

KW - keel

KW - laying hens

KW - machine learning

KW - radiograph

KW - tibia

UR - https://res.slu.se/id/publ/132952

U2 - 10.54612/a.q33hkb4fja

DO - 10.54612/a.q33hkb4fja

M3 - Doctoral thesis

SN - 978-91-8046-447-5

T3 - Acta Universitatis Agriculturae Sueciae

PB - Swedish University of Agricultural Sciences

ER -