Agricultural University of Athens
Genome-wide association analysis and gene network analysis for (re)production traits in commercial broilers
Abstract
dc:descriptionAim of the present PhD thesis was to identify genetic variants and plausible functional genes for two typical polygenic traits in broilers, one related to growth (body weight at 35 days of age, BW) and the other with reproduction efficiency (number of eggs per female broiler, EN). To this end, numerous genome-wide association studies (GWAS) were conducted to detect marker-trait associations and various post-GWAS in silico methods were applied in efforts to discover (novel) functional candidate genes for individual and joint traits.Chapter 1 presents an optimal sampling strategy when conducting cost-effective GWASs. To this end, 19 GWASs for BW were conducted using random and extreme phenotype (continuous and dichotomized) samples with sizes ranging from 5% to 50% of a total population of 6700 broilers. The most efficient sampling scenario was identified by comparing genome-wide significant marker signals between sub-samples and the whole population. This comparison pointed out 50% extreme phenotype sampling as the optimal sampling scenario. In the optimal sampling scenario, putative genetic variants were selected using a threshold of 1% for the Proportion of Variance Explained associated with markers. This search strategy resulted in identification of a total number of five putative causal genetic variants. These variants resided in genomic regions harboring ten growth-related QTLs (e.g. breast muscle percentage, abdominal fat weight etc.) and six growth related genes (CACNB1, MYOM2, SLC20A1, ANXA4, FBXO32, SLAIN2). Chapter findings proposed the use of 50% extreme phenotype sampling as the optimal sampling strategy to detect causative genes when performing cost-effective GWASs. Chapter 2 investigates the utility of various computational (in silico) techniques to propose functional candidate genes for BW. First, a GWAS for BW was conducted to detect significant marker-trait associations. This analysis pointed out twelve genome-wide significant SNPs across nine autosomes. The search for positional candidate genes in 1 Mb flanking regions around the significant markers revealed a total number of 1,012 candidate genes and 197 growth-related QTL/associations. Implementation of Functional Enrichment Analysis (FEA), Pathway Analysis (PA), Gene Network Analysis (GNA) and Gene functional Prioritization Analysis (GPA) to predict functional relevance of the candidate genes followed. FEA pointed out 49 candidate genes participating in system development while PA highlighted 25 member genes in growth-related pathways. GPA and GNA pointed out 14 common genes (UBC, SMAD4, SHC1, NRAS, PSMD4, CDC6, PSMD7, RARA, PSMB4, CDH1, STAT5B, MED1, PSMD3 and CDT1) with highest functional relevance to the trait under study. Application of the preceding computational methods has, individually or jointly, demonstrated that extant knowledge and available tools can be useful to prioritize most likely candidates for traits under investigation. In Chapter 3, the functional role of modular genes (genes organized into dense sub-networks) was explored in efforts to identify causal genes for BW. To this end, first a GWAS for BW was carried out. Then, genomic regions around the significant SNPs showing strong linkage disequilibrium levels (D′>0.8) were searched to identify positional candidate genes and reported QTLs for the trait. This search revealed a total number of 645 positional candidate genes around 11 genome-wide significant markers. Community structure analysis was then conducted to detect densely interconnected nodes (modules) of the positional candidate genes. This analysis detected five modules formed by 401 candidate genes. Functional enrichment analysis of the modular genes showed 52 genes as participating in developmental processes. 14 more modular genes (GABRG1, NGF, APOBEC2, STAT5B, STAT3, SMAD4, MED1, CACNB1, SLAIN2, LEMD2, ZC3H18, TMEM132D, FRYL and SGCB) had evidenced growth functional relevance. In the present Chapter, a total number of 66 functional candidate genes for BW were proposed, some of which were novel and some identified in previous studies. Chapter 4 aimed to provide the genetic variants impacting on egg number (EN) in female broilers. To this end, additive and dominant genetic models were applied to detect trait-marker associations. This analysis resulted in identification of a total number of 17 chromosome-wide significant markers. Of these, 7 were additive, 4 dominant and 6 additive plus dominant. Degree of dominance for the purely dominant markers ranged from partial to complete dominance. A total number of 57 positional candidate genes were identified within 50 kb flanking regions around the significant markers. Functional enrichment analysis of the positional candidate genes pinpointed two genes (BHLHE40 and CRTC1) to be involved in the ‘entrainment of circadian clock by photoperiod’ biological process. Gene prioritization analysis of the positional candidate genes identified ten top ranked genes (GDF15, BHLHE40, JUND, GDF3, COMP, ITPR1, ELF3, ELL, CRLF1 and IFI30). Seven prioritized genes (GDF15, BHLHE40, JUND, GDF3, COMP, ELF3, CRTC1) had documented functional relevance to reproduction, while two more prioritized genes (ITPR1 and ELL) are reported to be related to egg quality in female chickens.Chapter 5 aimed to shed light in the genetic mechanism underlying the antagonistic (negative) genetic correlation between growth (BW) and reproduction (EN) in broilers. First, a bivariate marker-traits association analysis was performed. This analysis resulted in detection of a total number of 51 genome-wide significant markers across 21 autosomes. Application of stepwise conditional-joint analyses pinpointed a total number of 13 independent markers. These independent markers were located within 17 reported QTLs and within or close proximity to 14 positional candidate genes. Examination of the Gene Ontology (GO) Biological Process (BP) profile per candidate gene highlighted two genes (ACVR1 and CACNA1H) that were participating in relevant BPs to the traits under study. Of these genes, ACVR1 was directly involved in both developmental and reproductive processes while CACNA1H was directly participating in reproduction and indirectly in growth via metabolic processes. Literature evidence of the functions of the two aforementioned genes (ACVR1 and CACNA1H) enhanced their candidacy as pleiotropic genes for the examined traits and set them as typical gene exemplars of horizontal pleiotropy. Chapter findings provide novel insight in the nature of the antagonistic co-variation between growth and reproduction in female broilers.
Degree
thesis:*- Grantor dc:publisher
- Agricultural University of Athens
- Year dc:date
- 2020
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Tarsani, Eirini
- Ταρσάνη, Ειρήνη
Subjects
dc:subject × 20- Κρεοπαραγωγικά ορνίθια
- Σωματικό βάρος στις 35 ημέρες
- Αριθμός αυγών ανά όρνιθα
- Γονιδιωματικής κλίμακας αναλύσεις συσχέτισης
- Βέλτιστη στρατηγική δειγματοληψίας
- Εργαλεία βιοπληροφορικής
- Επιδράσεις κυριαρχίας
- Πλειοτροπική δράση
- Broilers
- Body weight at 35 days of age
- Number of eggs per female broiler
- Genome-wide association studies
- Optimal sampling strategy
- Computational approaches
- Dominant genetic effects
- Pleiotropy
- Γεωπονικές Επιστήμες και Κτηνιατρική
- Επιστήμη Ζωικής Παραγωγής
- Agricultural and Veterinary Sciences
- Animal and Dairy Science
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/47810