Chicken genome variations and selection: from sequences to consequences

Abstract

Chicken is a major protein source and intensively selected for economically important traits by humans. As such, this generated a huge range of phenotypes that representing a diverse spectrum of genetic variation. Understanding the functional basis of the genetic variants that underlie these traits, however, remains a formidable endeavour particularly for complex traits. Nonetheless, molecular phenotyping of an organism from sequenced data is doable with the advances in bioinformatics analysis and unparalleled surveys of genome wide genetic variants. This provides the opportunity to gain insights into the genome architecture and assists in identifying chromosomal regions underlying selection through a “sequences to consequences” approach. Combining a whole genome re-sequencing (WGS) approach with the knowledge of selection history, this thesis aimed to study the chromosomal regions and genetic variants underlying traits of interest in various selected chicken populations. To achieve this, genetic (quantitative and population genetics), genomic and bioinformatics approaches were employed and integrated to investigate the genome wide selection signatures in a number of different lines of chicken selected for different complex traits. This includes analysing: (i) divergently selected broilers for fatness traits (Chapter 2), (ii) a closed population of layer chickens (Chapter 3), (iii) selection signatures unique to broiler or layer chickens (Chapter 4) and (iv) selection signatures in colony stimulating factor 1 (CSF1) associated with gene expression differences in broiler and layer populations (Chapter 5). Candidate genes and nucleotides underlying potential selection regions were identified, and attempts were made to further elucidate the potential interplay between genes and the biological pathways involved in regulating traits in these selected chicken lines. Incorporating integrative approaches, variants within selection signatures were annotated to provide further evidence of their functional consequences. Overall, non-coding regions were enriched in selection signatures implied that causative variants may have regulatory roles. Capitalising on the millions of genetic variants discovered from WGS, chromosomal regions subject to selection were detected using a number of population genetics statistics. In broiler chicken lines divergently selected for very low-density plasma lipoprotein (VLDL) (Chapter 2), incorporating signatures of selection helped to improve the resolution of previously mapped quantitative traits loci (QTL) intervals. This research demonstrated that the integration of the analysis of selection signatures with functional annotation of genetic variants enabled refinement and characterisation of the QTL for fatness traits. In a closed population of brown leghorn layers (Chapter 3), evidence of selection signatures was found through Tajima’s D analysis. The analysis unravelled selection signatures encoding genes involved in numerous pathways and genes having key roles such as in behaviour, including feather pecking. Combining population differentiation statistic (FST) and Tajima’s D, a number of regions subject to divergent selection between broilers and white egg layers were identified (Chapter 4). Selection signatures were found to harbour mutations involved in cellular and tissue development, including genes having important roles in growth, fatness, egg shell strength and muscle development. These regions and the overlapping genes thereby may be potentially contributing to the different phenotypic variations observed between broilers and layers. In Chapter 5, a revised gene model for colony stimulating factor 1 (CSF1) showing complex pattern of alternate transcripts was predicted from transcriptome analysis of RNA isolated from 21 different tissues. In parallel, selection signatures analysis with the FST statistic, identified selection signatures that differentiate broilers to white egg layers (3 regions) or brown egg layers to white egg layers (4 regions). All these selection signatures were located within non-coding regions, indicating potential divergent selection of CSF1 within regulatory regions. Overall, the results presented in this thesis using the “sequences to consequences” approach, link several genomic regions and genes to phenotypic variation in domesticated chicken lines. The work reported here serves as a foundation for further study to decipher the relationship between “genotype and phenotype” and its functional consequences due to selection.