Genome-wide gene expression surveys and a transcriptome map in chicken

The chicken (Gallus gallus) is an important model organism in genetics, developmental biology, immunology, evolutionary research, and agricultural science. The completeness of the draft chicken genome sequence provided new possibilities to study genomic changes during evolution by comparing the chicken genome to that of other species. The development of long oligonucleotide microarrays based on the genome sequence made it possible to survey genome-wide gene expression in chicken. This thesis describes two gene expression surveys across a range of healthy chicken tissues in both adult and embryonic stages. Specifically, we focus on the mechanisms of regulation of gene transcription and their evolution in the vertebrate genome. Chapter 1 provides a brief history of the chicken as a model organism in biological and genomics research. In particular a brief overview is presented about expression profiling experiments, followed by an introduction to gene transcription regulation in general. Finally, the aim and outline of this thesis is presented. An important aim of this thesis is to generate surveys of genome-wide gene expression data in chicken using microarrays. In chapter 2, we introduce microarray data normalization including background correction, within-array normalization and between-array normalization. Based on these results an analysis approach is recommended for the analysis of two-color microarray data as performed in the experiments described in this thesis. We also briefly explain the relevant methodology for the identification of differentially expressed genes and how to translate resulting gene lists into biological knowledge. Finally, specific issues related to updating microarray probe annotation in farm animals, is discussed. For the analysis of the microarray data in this thesis re-annotation of the probes on the chicken 20K oligoarray was done using the oligoRAP, analysis pipeline. The vast amount of data generated from a single transcriptomics study makes it impossible to extract meaningful biological knowledge by manually going through individual genes from a list with hundreds and thousands of differentially expressed genes. In chapter 3, we present a practical approach using a collection of R/Bioconductor packages to extract biological knowledge from a microarray experiment in farm animals. Furthermore, a locally adaptive statistical procedure (LAP) analysis approach is used to identify differentially expressed chromosomal regions in a microarray experiment. Chapter 4 presents a genome-wide gene expression survey across eight different tissues (brain, bursa of Fabricius, kidney, liver, lung, small intestine, spleen, and thymus from 10-week old chickens) in adult birds using a chicken 20K microarray. To a certain extent, most genes show some tissue-specific pattern of expression. Housekeeping and tissue-specific genes are identified based on gene expression patterns across the eight different tissues. The results show that housekeeping genes are more compact, i.e. are smaller, with shorter, coding sequence length, intron length, and smaller length of the intergenic regions. This observed compactness of housekeeping genes may be a result of selection on economy of transcription during evolution. Furthermore, a comparative analysis of gene expression among mouse, chicken, and frog showed that the expression patterns of orthologous genes are conserved during evolution between mammals, birds, and amphibians. The chicken embryo has been a very popular model for developmental biology. To study the overall gene expression pattern in whole chicken embryos at different developmental stages and/or embryonic tissues, a genome-wide gene expression survey across different developmental and embryonic stages was performed (chapter 5). The study included four different developmental stages (HH stage 3, 10, 15, 22) and eight different embryonic tissues (brain, bursa of Fabricius, heart, kidney, liver, lung, small intestine, and spleen from HH stage 36). We were able to identify several embryonic stage- and tissue-specific genes in our analysis. Genomic features of genes widely expressed under these 12 conditions suggest that widely expressed genes are more compact than tissue-specific genes, confirming the findings described in chapter 4. The analysis of the differentially expressed genes during the different developmental stages of whole embryo indicates a gradual change in gene expression during embryo development. A comparison of the gene expression profiles between the same organs, of adults and embryos reveals both striking similarities as well as differences. The overall goal of this thesis was to improve our understanding of the mechanisms of transcriptional regulation in the chicken. In chapter 6, a transcriptome map for all chicken chromosomes is presented based on the expression data described in chapter 4. The results reveal the presence of two distinct types of chromosomal regions characterized by clusters of highly or lowly expressed genes respectively. Furthermore, these regions show a high correlation with a number of genome characteristics, like gene density, gene length, intron length, and GC content. A comparative analysis between the chicken and human transcriptome maps suggests that the regions with clusters of highly expressed genes are relatively conserved between the two genomes. Our results revealed the presence of a higher order organization of the chicken genome that affects gene expression, confirming similar observations in other species. Finally, in chapter 7 I summarize the main findings and discuss some of the limitations of the analyses described in this thesis. I also discuss the different merits and shortcomings of studying gene expression using either microarrays or next-generation sequencing technology and propose directions for future research. The rapid developments in new-generation sequencing technology will facilitate better coverage and depth of the chicken genome. This will provide a better genome assembly and an improved genome annotation. The sequence-based approaches for studying gene expression will reduce noise levels compared to hybridization-based approaches. Overall, next-generation sequencing is already providing greatly enhance tools to further improve our understanding of the chicken transcriptome and its regulation.