Search

Phosphorylation of caseins (CN) is a crucial post-translational modification allowing caseins to aggregate as micelles. The formation and stability of casein micelles are important for transporting abundant minerals to the neonate and manufacturing of dairy products. Therefore, it is of great interest to explore variation in degrees of phosphorylation of caseins and study to what extent genetic and other factors contribute to this variation. This thesis aimed to investigate the genetic background of bovine milk protein composition with a focus on phosphorylation of αs1- and αs2-CN. Thus, two studies were conducted to quantify phosphorylation levels of αs1- and αs2-CN: one in French Montbéliarde (FM) using liquid chromatography coupled with electrospray ionization mass spectrometry and the other in Dutch Holstein Friesian (DHF) using capillary zone electrophoresis. In FM, in addition to the known isoforms αs1-CN-8P and-9P and αs2-CN-10P to -13P, three new phosphorylation isoforms were detected, namely αs2-CN-9P, αs2-CN-14P, and αs2-CN-15P. Relative concentrations of the phosphorylation isoforms varied considerably among cows. Phenotypic correlations showed that isoforms phosphorylated at higher degrees (αs1-CN-9P and αs2-CN-12P to -14P) correlated negatively with isoforms phosphorylated at lower degrees (αs1-CN-8P, αs2-CN-10P, and -11P). Furthermore, it was shown that αs1- and αs2-CN phosphorylation profiles changed across parity and lactation, and exploitable genetic variation for the phosphorylation degrees of αs1- and αs2-CN (defined as the proportion of higher-degree isoforms in αs1- and αs2-CN, respectively) exists in FM. In DHF, three αs2-CN isoforms, namely αs2-CN-10P to -12P, and the phosphorylation degrees of αs1- and αs2-CN were quantified. High intra-herd heritabilities were estimated for individual αs2-CN phosphorylation isoforms and the phosphorylation degrees of αs1- and αs2-CN (ranging from 0.54 to 0.89). This suggests that genetic factors contribute substantially to observed differences in αs1- and αs2-CN phosphorylation profiles. The correlation between the phosphorylation degrees of αs1- and αs2-CN was 0.94. Additionally, a total of 10 regions, distributed across Bos taurus autosomes (BTA) 1, 2, 6, 9, 11, 14, 15, 18, 24 and 28, were detected to be associated with individual αs1- and αs2-CN phosphorylation isoforms and their phosphorylation degrees in DHF. Regions on BTA1, 6, 11 and 14 were associated with multiple traits studied. Two quantitative trait loci (QTL) regions were detected on BTA1: one affecting αs2-CN production, and the other affecting αs1-CN PD and αs2-CN PD. The QTL region on BTA6 affected only individual αs2-CN isoforms. The QTL region on BTA11 and 14 affected relative concentrations of αs2-CN-10P and αs2-CN-11P, αs1-CN PD and αs2-CN PD. Results suggested that effects of identified genomic regions on αs1-CN PD and αs2-CN PD are probably due to changes in milk synthesis and phosphorus secretion in milk. Finally, differences among studies due to factors such as analytical methods, trait definitions, and breed on genetic parameters and correlations are discussed using the two dataset from this thesis. It is concluded that differences in heritability estimates for αs1-CN-8P and -9P, αs2-CN-10P, -11P and -12P, and αs1-CN PD between FM and DHF were mainly due to genetic differences between breeds. As for αs2-CN PD, it was defined differently in FM and DHF due to analytical methods used. It is shown that both trait definitions successfully quantified the proportion of isoforms with higher degrees of phosphorylation because of similar estimated correlations using both definitions on the FM dataset. Additionally, it is hypothesized that a two-casein-kinase system is involved in the phosphorylation of αs1- and αs2-CN based on results in this thesis.; La phosphorylation des caséines (CN) est une modification post-traductionnelle qui permet l’aggrégation des caséines sous formes de micelles. La stabilité de ces structures est essentielle pour transporter les minéraux au nouveau-né et fabriquer des produits laitiers. Il est par conséquent fondamental d'explorer la variation du degré de phosphorylation (PD) des CN, et d'étudier dans quelle mesure les facteurs génétiques contribuent à cette variation. Cette thèse visait à étudier le contexte génétique de la composition des protéines du lait de vache, en mettant l'accent sur la phosphorylation des CN αs1 et αs2. Deux études ont été menées pour analyser leur niveau de phosphorylation : la première en race Montbéliarde française (FM), en utilisant la chromatographie liquide couplée à la spectrométrie de masse à ionisation par électro-nébulisation et l'autre en race Holstein Friesian néerlandaise (DHF), en utilisant l’électrophorèse capillaire de zone. En race FM, en plus des isoformes connues : αs1-CN-8P et -9P, de αs2-CN-10P à -13P, trois nouvelles isoformes ont été détectées (αs2-CN-9P, αs2-CN-14P et αs2-CN-15P). Les concentrations relatives de ces isoformes variaient considérablement entre vaches. Les corrélations phénotypiques ont montré que les isoformes phosphorylées aux degrés les plus élevés (αs1-CN-9P, de αs2-CN-12P à -14P) étaient corrélées négativement avec les isoformes phosphorylées à des degrés inférieurs (αs1-CN-8P, αs2-CN-10P et -11P). En outre, les profils de phosphorylation des CN αs1 et αs2 changent avec la parité et le stade de lactation, et il est possible d’exploiter en FM la variation génétique du PD de ces caséines (définie comme la proportion d'isoformes ayant les taux de phosphorylation les plus élevés). En DHF, nous avons quantifié 3 isoformes d’αs2-CN, à savoir αs2-CN-10P, -11P, et -12P et les PD des CN αs1 et αs2. Des héritabilités intra-troupeau élevées ont été estimées pour les isoformes de phosphorylation d'αs2-CN et les PD des CN αs1 et αs2 (comprise entre 0,54 et 0,89). Ceci suggère que des facteurs génétiques contribuent substantiellement aux différences observées dans les profils de phosphorylation de ces CN. La correlation entre les degrés de phosphorylation de αs1-CN et d’αs2-CN était de 0,94. Nous avons montré que 10 régions, réparties sur les autosomes BTA1, 2, 6, 9, 11, 14, 15, 18, 24 et 28, sont associées aux isoformes de phosphorylation des CN αs1 et αs2 et leurs PD en DHF. Les régions sur BTA1, 6, 11 et 14 étaient associées à plusieurs caractères étudiés. Deux régions quantitative trait loci (QTL) ont été détectées sur BTA1 : l'une affectant la production d'αs2-CN et l'autre le PD des deux CN. La région QTL localisée sur BTA6 affectait uniquement les isoformes d'αs2-CN. Les régions QTL situées sur BTA11 et BTA14 influencent les concentrations relatives en αs2-CN-10P et -11P, et les PD des CN αs1 et αs2. Ces résultats suggèrent que les effets des régions génomiques identifiées impactant les PD sont probablement dus à des modifications de biosynthèse des constituants du lait et à la sécrétion de phosphore dans le lait. Enfin, les différences observées sur les paramètres génétiques estimés et les corrélations, sont discutées sur la base de l'ensemble des données issues des deux études décrites dans cette thèse. En conclusion, les différences observées sur les estimations d'héritabilité entre FM et DHF semblent principalement dues à des composantes génétiques. Quant au PD de l'αs2-CN, il a été défini différemment en FM et en DHF, en raison des méthodes analytiques. Il a été possible de quantifier avec succès la proportion d'isoformes présentant des degrés de phosphorylation élevés, en raison de la similitude dans les corrélations estimées avec les deux définitions aux données de FM. Finalement, sur la base des résultats présentés dans cette thèse, nous avons émis l'hypothèse qu'un système comportant deux caséines-kinases puisse être impliqué dans la phosphorylation des CN αs1 et αs2.

Phosphorylation of caseins (CN) is a crucial post-translational modification allowing caseins to aggregate as micelles. The formation and stability of casein micelles are important for transporting abundant minerals to the neonate and manufacturing of dairy products. Therefore, it is of great interest to explore variation in degrees of phosphorylation of caseins and study to what extent genetic and other factors contribute to this variation. This thesis aimed to investigate the genetic background of bovine milk protein composition with a focus on phosphorylation of αs1- and αs2-CN. Two studies were conducted to quantify phosphorylation levels of αs1- and αs2-CN: one in French Montbéliarde using liquid chromatography coupled with electrospray ionization mass spectrometry and the other in Dutch Holstein Friesian using capillary zone electrophoresis. In French Montbéliarde, in addition to the known isoforms αs1-CN-8P and-9P and αs2-CN-10P to -13P, three new phosphorylation isoforms were detected, namely αs2-CN-9P, αs2-CN-14P, and αs2-CN-15P. Relative concentrations of the phosphorylation isoforms varied considerably among cows. Phenotypic correlations showed that isoforms phosphorylated at higher degrees (αs1-CN-9P and αs2-CN-12P to -14P) correlated negatively with isoforms phosphorylated at lower degrees (αs1-CN-8P, αs2-CN-10P, and -11P). Furthermore, it was shown that αs1- and αs2-CN phosphorylation profiles changed across parity and lactation, and exploitable genetic variation for the phosphorylation degrees of αs1- and αs2-CN (defined as the proportion of higher-degree isoforms in αs1- and αs2-CN, respectively) exist. In Dutch Holstein Friesian, three αs2-CN isoforms, namely αs2-CN-10P to -12P, and the phosphorylation degrees of αs1- and αs2-CN were quantified. High intra-herd heritabilities were estimated for individual αs2-CN phosphorylation isoforms and the phosphorylation degrees of αs1- and αs2-CN (ranging from 0.54 to 0.89). This suggests that genetic factors contribute substantially to observed differences in αs1- and αs2-CN phosphorylation profiles. The highly positive correlation between the phosphorylation degrees of αs1- and αs2-CN (0.94) suggest that phosphorylation of αs1- and αs2-CN is related. Additionally, a total of 10 regions, distributed across Bos taurus autosomes (BTA) 1, 2, 6, 9, 11, 14, 15, 18, 24 and 28, were detected to be associated with individual αs1- and αs2-CN phosphorylation isoforms and their phosphorylation degrees. Regions on BTA1, 6, 11 and 14 were associated with multiple traits studied. Two quantitative trait loci (QTL) regions were detected on BTA1: one affecting αs2-CN production, and the other affecting αs1-CN PD and αs2-CN PD. The QTL region on BTA6 affected only individual αs2-CN isoforms. The QTL region on BTA11 and 14 affected relative concentrations of αs2-CN-10P and αs2-CN-11P, αs1-CN PD and αs2-CN PD. Results suggested that effects of identified genomic regions on αs1-CN PD and αs2-CN PD are probably due to changes in milk synthesis and phosphorus secretion in milk.

Duchemin, S.I. (2016). Mapping and fine-mapping of genetic factors affecting bovine milk composition. Joint PhD thesis, between Swedish University of Agricultural Sciences, Sweden and Wageningen University, the Netherlands Bovine milk is an important source of nutrients in Western diets. Unraveling the genetic background of bovine milk composition by finding genes associated with milk-fat composition and non-coagulation of milk were the main goals of this thesis. In Chapter 1, a brief description of phenotypes and genotypes used throughout the thesis is given. In Chapter 2, I calculated the genetic parameters for winter and summer milk-fat composition from ~2,000 Holstein-Friesian cows, and concluded that most of the fatty acids (FA) can be treated as genetically the same trait. The main differences between milk-fat composition between winter and summer milk samples are most likely due to differences in diets. In Chapter 3, I performed genome-wide association studies (GWAS) with imputed 777,000 single nucleotide polymorphism (SNP) genotypes. I targeted a quantitative trait locus (QTL) region on Bos taurus autosome (BTA) 17 previously identified with 50,000 SNP genotypes, and identified a region covering 5 mega-base pairs on BTA17 that explained a large proportion of the genetic variation in de novo synthesized milk FA. In Chapter 4, the availability of whole-genome sequences of keys ancestors of our population of cows allowed to fine-mapped BTA17 with imputed sequences. The resolution of the 5 mega base-pairs region substantially improved, which allowed the identification of the LA ribonucleoprotein domain family, member 1B (LARP1B) gene as the most likely candidate gene associated with de novo synthesized milk FA on BTA17. The LARP1B gene has not been associated with milk-fat composition before. In Chapter 5, I explored the genetic background of non-coagulation of bovine milk. I performed a GWAS with 777,000 SNP genotypes in 382 Swedish Red cows, and identified a region covering 7 mega base-pairs on BTA18 strongly associated with non-coagulation of milk. This region was further characterized by means of fine-mapping with imputed sequences. In addition, haplotypes were built, genetically differentiated by means of a phylogenetic tree, and tested in phenotype-genotype association studies. As a result, I identified the vacuolar protein sorting 35 homolog, mRNA (VPS35) gene, as candidate. The VPS35 gene has not been associated to milk composition before. In Chapter 6, the general discussion is presented. I start discussing the challenges with respect to high-density genotypes for gene discovery, and I continue discussing future possibilities to expand gene discovery studies, with which I propose some alternatives to identify causal variants underlying complex traits in cattle.

Background: Investigating the manner in which genetic and environmental factors interact to increase susceptibility to iron deficiency, has the potential to impact on strategies to overcome iron deficiency as well as the development of biomarkers to monitor iron status in populations. Single nucleotide polymorphisms or genetic variants that may affect the composition and hence the functionality of proteins involved in iron metabolism have been the subject of recent genetic association studies. However, these investigations have not yet been carried out in African populations that differ genetically from populations of European ancestry and which bear the highest burden of iron deficiency. The overall aim of this thesis was to investigate the genetics of iron status in African populations using a candidate gene approach. Methods: In order to evaluate the association between identified TMPRSS6 gene variants and iron status we conducted a systematic review with meta-analyses. We primarily searched the literature using the HuGE Navigator, Pubmed and Scopus databases for primarily genome wide association studies. Fixed effects meta-analysis was used to obtain summary estimates. Associations between reported variants and iron status as well as gene-gene and variant interactions that influence iron status were investigated in a female black South African cohort (n=686; range 32–86 years) which were part of the Prospective Urban and Rural Epidemiology (PURE) study. Concentrations of haemoglobin, serum ferritin, serum transferrin receptor and body iron stores were determined. Thirty SNPs were genotyped and passed all quality criteria. To investigate whether previously identified associations in populations of European ancestry are replicated in populations of African ancestry, we conducted candidate gene association studies. Twenty iron status-associated variants in 628 Kenyans, 609 Tanzanians, 608 South Africans and 228 African Americans were genotyped and associations investigated using haemoglobin and serum ferritin as outcome measures. Finally, we assessed the effect of TNF-α allele variants (TNF‒1031, TNF‒308) on malaria rates, the severity of malaria as indicated by haemoglobin concentrations at the time of presentation with febrile episodes and the association between Plasmodium infection and haemoglobin concentration in symptomless parasite carriers. We used data from a placebo-controlled trial which consisted of 612 Tanzanian children aged 6–60 months. Cox regression models were used in the primary analysis to account for multiple episodes per child. Results: In our systematic review we included eleven studies on Caucasian populations, four on Asian populations and one study on an African-American population. Differences in minor allele frequencies (MAF) of 8 TMPRSS6 SNPs (rs855791, rs4820268, rs2111833, rs1421312, rs228921, rs228918, rs228919 and rs575620) across ethnic groups were observed; with the MAF of rs855791 being significantly higher in Asian populations than in Caucasians (0.55 vs 0.42). In the meta-analysis, the A allele of rs855791 was associated with lower haemoglobin and ferritin concentrations in all populations. This allele was also associated with increased serum transferrin receptor and transferrin concentrations. We observed similar associations for the G allele in rs4820268. In general, minor allele frequencies (MAF) from females in the PURE population were lower compared to those of males and females of European ancestry populations in the 1000 Genomes Project. In the TF gene, the SNP rs1799852 was associated with decreased serum ferritin (p=0.01) and body iron concentrations (p=0.03) and increased serum transferrin receptor (sTfR) concentrations (P=0.004), while rs3811647 was associated with transferrin receptor and body iron (both P=0.03) in a U-shaped manner. The chromosome 6 SNP allele combination (AAA) consisting of rs1799964 and rs1800629 both in TNF α and rs2071592 in NFKBIL1 was associated with higher odds for low serum ferritin concentrations (serum ferritin8.3mg/L; OR:0.79 (95%-CI,0.63-0.98). We successfully replicated reported significant associations with lowered haemoglobin concentrations for two loci in TMPRSS6 namely rs2413450 and rs4820268 and with increased haemoglobin concentrations for one locus in TF (rs3811658) when analysing the four populations of African ancestry. When ferritin was considered as an outcome measure, we replicated associations with increased ferritin concentrations in two loci namely, rs228918 in TMPRSS6 and rs1525892 in TF. No other significant associations were determined. Malaria rates were higher in Tanzanian children with the TNF‒1031CC genotype (rs1799964) compared to the AA genotype (crude hazard ratio (HR), 95%CI: 1.41 [1.01‒1.97], adjusted HR 1.31 [0.97‒1.76]y) but were lower in those with the TNF‒308AA genotype (rs1800629) (adjusted HR 0.13 [0.02‒0.63]) compared to those harbouring the wild type homozygous genotype. Conclusions: This thesis demonstrates previously observed associations between TMPRSS6 gene variants and haemoglobin concentrations in European ancestry populations are replicated in African populations. Replication of results in other loci previously associated with iron status in European ancestry populations was not achieved. Additionally, minor allele frequencies of single nucleotide polymorphisms associated with iron status are generally higher in European ancestry cohorts compared to those of African ancestry populations. The lack of association of reported variants may indicate that novel loci are responsible for the heritability of iron status in African populations. We have additionally observed that TNF α variants increase malaria severity. Malaria is a major cause of iron deficiency in malaria endemic areas. Our finding emphasizes that to alleviate iron deficiency in malaria endemic areas prevention and treatment of malaria is necessary. This thesis highlights the need to conduct genetic association studies in African populations where iron deficiency is of utmost public health significance. In addition, investigations into the genetics of iron status are bound to contribute towards the development of biomarkers that are useful in the determination of iron status in areas of high inflammation burden.

Background: Investigating the manner in which genetic and environmental factors interact to increase susceptibility to iron deficiency, has the potential to impact on strategies to overcome iron deficiency as well as the development of biomarkers to monitor iron status in populations. Single nucleotide polymorphisms or genetic variants that may affect the composition and hence the functionality of proteins involved in iron metabolism have been the subject of recent genetic association studies. However, these investigations have not yet been carried out in African populations that differ genetically from populations of European ancestry and which bear the highest burden of iron deficiency. The overall aim of this thesis was to investigate the genetics of iron status in African populations using a candidate gene approach. Methods: In order to evaluate the association between identified TMPRSS6 gene variants and iron status we conducted a systematic review with meta-analyses. We primarily searched the literature using the HuGE Navigator, Pubmed and Scopus databases for primarily genome wide association studies. Fixed effects meta-analysis was used to obtain summary estimates. Associations between reported variants and iron status as well as gene-gene and variant interactions that influence iron status were investigated in a female black South African cohort (n=686; range 32–86 years) which were part of the Prospective Urban and Rural Epidemiology (PURE) study. Concentrations of haemoglobin, serum ferritin, serum transferrin receptor and body iron stores were determined. Thirty SNPs were genotyped and passed all quality criteria. To investigate whether previously identified associations in populations of European ancestry are replicated in populations of African ancestry, we conducted candidate gene association studies. Twenty iron status-associated variants in 628 Kenyans, 609 Tanzanians, 608 South Africans and 228 African Americans were genotyped and associations investigated using haemoglobin and serum ferritin as outcome measures. Finally, we assessed the effect of TNF-α allele variants (TNF‒1031, TNF‒308) on malaria rates, the severity of malaria as indicated by haemoglobin concentrations at the time of presentation with febrile episodes and the association between Plasmodium infection and haemoglobin concentration in symptomless parasite carriers. We used data from a placebo-controlled trial which consisted of 612 Tanzanian children aged 6–60 months. Cox regression models were used in the primary analysis to account for multiple episodes per child. Results: In our systematic review we included eleven studies on Caucasian populations, four on Asian populations and one study on an African-American population. Differences in minor allele frequencies (MAF) of 8 TMPRSS6 SNPs (rs855791, rs4820268, rs2111833, rs1421312, rs228921, rs228918, rs228919 and rs575620) across ethnic groups were observed; with the MAF of rs855791 being significantly higher in Asian populations than in Caucasians (0.55 vs 0.42). In the meta-analysis, the A allele of rs855791 was associated with lower haemoglobin and ferritin concentrations in all populations. This allele was also associated with increased serum transferrin receptor and transferrin concentrations. We observed similar associations for the G allele in rs4820268. In general, minor allele frequencies (MAF) from females in the PURE population were lower compared to those of males and females of European ancestry populations in the 1000 Genomes Project. In the TF gene, the SNP rs1799852 was associated with decreased serum ferritin (p=0.01) and body iron concentrations (p=0.03) and increased serum transferrin receptor (sTfR) concentrations (P=0.004), while rs3811647 was associated with transferrin receptor and body iron (both P=0.03) in a U-shaped manner. The chromosome 6 SNP allele combination (AAA) consisting of rs1799964 and rs1800629 both in TNF α and rs2071592 in NFKBIL1 was associated with higher odds for low serum ferritin concentrations (serum ferritin8.3mg/L; OR:0.79 (95%-CI,0.63-0.98). We successfully replicated reported significant associations with lowered haemoglobin concentrations for two loci in TMPRSS6 namely rs2413450 and rs4820268 and with increased haemoglobin concentrations for one locus in TF (rs3811658) when analysing the four populations of African ancestry. When ferritin was considered as an outcome measure, we replicated associations with increased ferritin concentrations in two loci namely, rs228918 in TMPRSS6 and rs1525892 in TF. No other significant associations were determined. Malaria rates were higher in Tanzanian children with the TNF‒1031CC genotype (rs1799964) compared to the AA genotype (crude hazard ratio (HR), 95%CI: 1.41 [1.01‒1.97], adjusted HR 1.31 [0.97‒1.76]y) but were lower in those with the TNF‒308AA genotype (rs1800629) (adjusted HR 0.13 [0.02‒0.63]) compared to those harbouring the wild type homozygous genotype. Conclusions: This thesis demonstrates previously observed associations between TMPRSS6 gene variants and haemoglobin concentrations in European ancestry populations are replicated in African populations. Replication of results in other loci previously associated with iron status in European ancestry populations was not achieved. Additionally, minor allele frequencies of single nucleotide polymorphisms associated with iron status are generally higher in European ancestry cohorts compared to those of African ancestry populations. The lack of association of reported variants may indicate that novel loci are responsible for the heritability of iron status in African populations. We have additionally observed that TNF α variants increase malaria severity. Malaria is a major cause of iron deficiency in malaria endemic areas. Our finding emphasizes that to alleviate iron deficiency in malaria endemic areas prevention and treatment of malaria is necessary. This thesis highlights the need to conduct genetic association studies in African populations where iron deficiency is of utmost public health significance. In addition, investigations into the genetics of iron status are bound to contribute towards the development of biomarkers that are useful in the determination of iron status in areas of high inflammation burden.

Dit rapport beschrijft de opbouw en en het gebruik van de biobank van de RIVM-centra CVG en PZO. Deze biobank is voor het grootste deel opgebouwd tussen 1987 en 1998 in het kader van drie grootschalige epidemiologische projecten. De deelnemers (ca. 60 000) aan deze projecten hebben bloed afgestaan, dat is opgeslagen in vriezers van -20 graden Celsius, -86 graden Celsius en -196 graden Celsius. Het bloed is verdeeld in fracties (plasma, serum, witte bloedcellen, rode bloedcellen) en opgeslagen in porties van 0,5 ml en 1,5 ml. Het bloed is al voor verschillende onderzoeksvraagstellingen op gebied van chronische ziekten gebruikt, waarbij het bloed van het oudste project het meest is uitgeput. Ook is er DNA opgewerkt. Om bloed veilig te stellen voor onderzoek naar specifieke chronische ziekten of andere kenmerken zijn er zogenaamde beschermde cohorten ingesteld. Op dit moment zijn er vier beschermde cohorten: hart- en vaatziekten, kanker, migraine, en overleden deelnemers.

The ability to learn and form memory has been demonstrated in various animal species, ranging from relatively simple invertebrates, such as snails and insects, to more complex vertebrate species, including birds and mammals. The opportunity to acquire new skills or to adapt behaviour through learning is an obvious benefit. However, memory formation is also costly: it can be maladaptive when unreliable associations are formed and the process of memory formation can be energetically costly. The balance between costs and benefits determines if learning and memory formation are beneficial to an animal or not. Variation in learning abilities and memory formation between species is thought to reflect species-specific differences in ecology. This thesis focused on variation in the number of trials required to form long-term memory (LTM). LTM is considered the most stable and durable type of memory, but also the most costly, because it requires protein synthesis. Many animal species require multiple learning experiences, which are spaced in time, to form LTM. This allows re-evaluation of information before an animal invests in costly LTM. There is, however, variation in the number of trials that animal species require to induce LTM formation. A number of insect species, including a number of parasitic wasp species, form LTM after only a single learning experience. Parasitic wasps can learn odours that guide them towards suitable hosts for their offspring, so-called oviposition learning. Substantial differences in LTM formation are observed among closely related species of parasitic wasps, which provides excellent opportunities for comparative studies. Both ecological and genetic factors involved in variation in LTM formation have been studied in this project. A multidisciplinary approach is essential to understand the evolution of variation in LTM formation, because the interaction between genes and environment shapes learning and memory formation. LTM formation was studied in closely related species of the genus Nasonia. These small parasitic wasps (~2 mm in length) lay their eggs in various species of fly pupae and differences in the ecology of the four known species of this genus have been described. A high-throughput method for olfactory conditioning was developed in which the wasps associated an odour, either chocolate or vanilla, with the reward of a host. A T-maze olfactometer was designed for high-throughput testing of memory retention. Using these methods, variation in memory retention was observed between three Nasonia species. Both N. vitripennis and N. longicornis form a long-lasting memory after a single conditioning trial, which lasts at least 5 days. Nasonia giraulti, on the other hand, lost its memory after 1 to 2 days after a single conditioning trial. Further studies focused on the difference between N. vitripennis and N. giraulti, which was most pronounced. By inhibiting LTM with transcription and translation inhibitors, it was confirmed that N. vitripennis forms this type of memory after a single conditioning trial. LTM is visible 4 days after conditioning in N. vitripennis. Nasonia giraulti does not form LTM after a single conditioning trial. Long-lasting memory is only formed after two trials, with a 4-hour interval between them. This difference in LTM formation makes N. vitripennis and N. giraulti excellent model species to study both ecological and genetic factors involved in this difference. Ecological factors such as the value of the reward and the reliability of the learned association have been shown to affect memory formation in a number of animal species. A recent study on oviposition learning in two parasitic wasp species demonstrated that LTM formation depends on the host species, i.e. the reward offered during conditioning. LTM was formed when a host with a higher quality was offered, but not when a host of lower quality was offered. The effect of host quality on memory retention of N. vitripennis and N. giraulti was tested. Either a large host, Calliphora vomitoria, a medium-sized host, Lucilia sericata, or a small host, Musca domestica, was offered during conditioning. These hosts were observed to differ significantly in their quality, i.e. in the number of parasitoid offspring that emerged and the size of the offspring. There was, however, no effect of host species on memory retention in either Nasonia species. These results suggest that host quality is not important for LTM formation in N. vitripennis and N. giraulti. This observation shows that ecological factors that are important for memory formation in one species may not be important for another species. The genetic basis of memory formation is highly conserved among distant animal phyla. A large number of genes involved in LTM formation have been identified in genetic model organisms, including fruit flies, honeybees, the California sea hare, mice and rats, and the zebra finch.Genetic factors responsible for natural variation in LTM formation between species are currently unknown, however. Two approaches were used to study genetic factors responsible for the difference in LTM formation between N. vitripennis and N. giraulti. The first approach took advantage of the unique possibility to interbreed Nasonia species. Hybrid offspring of N. vitripennis and N. giraulti did not form LTM after a single conditioning trial, similar to N. giraulti. The dominant LTM phenotype of N. giraulti was then backcrossed into the genetic background of N. vitripennis for up to 5 generations. Using a genotyping microarray analysis and subsequent confirmation experiments, we detected two genomic regions (quantitative trait loci – QTLs) that both reduce long-lasting memory, but not completely remove this memory. These results indicate that multiple QTLs regulate the difference in LTM formation between the two Nasonia species. Concluding, our approach has provided insights in the genomic basis of a naturally occurring difference in LTM formation between two species. Excellent opportunities for fine-scale QTL mapping are available for the genus Nasonia. This will allow identification of decisive regulatory mechanisms involved in LTM formation that are located in the two genomic regions detected in this study. The second approach took advantage of next-generation sequencing techniques that allow transcriptome-wide studies of gene expression levels. RNA from heads of N. vitripennis and N. giraulti was collected before conditioning and immediately, 4 hours, or 24 hours after conditioning. This RNA was sequenced strand-specifically using HiSeq technology, which allows detection of sense and antisense transcripts. Various genes, from a number of different signalling pathways known to be involved in LTM formation, were uniquely differentially expressed after conditioning in N. vitripennis. These genes are likely involved in the ongoing process of LTM formation in this species. A number of other genes with a known role in LTM formation,including genes involved in dopamine synthesis and in the Ras-MAPK and PI3K signalling pathways, were uniquely differentially expressed in N. giraulti. These genes may have a role in a LTM inhibitory mechanism in this species. Antisense transcripts were detected for a number of known memory genes, which may indicate a role inregulation of transcription, alternative splicing, or translation. This study is the first to compare gene expression patterns after conditioning between two species that differ in LTM formation. The results provide promising candidate genes for future studies in which the regulation of these genes, the function of specific splice variants, and spatial expression patterns in the brain should be studied to understand how these genes are involved in the regulation of LTM formation. Learning and memory formation have an important role in animal and human behaviour.Novel and valuable insights on both ecological and genetic factors responsible for variation in LTM formation have been revealed by the research presented in this thesis. Integrating ecological factors and genetic factors is essential, as genes are the level on which ecological factors can drive the evolution of variation in learning and memory formation. The genus Nasonia has offered excellent opportunities for ecological research as well as unique opportunities for studies on genomic and genetic factors, which were addressed by comparing closely related species that differ in memory formation. This thesis provides the basis for the identification of genomic differences responsible for the difference in memory formation between Nasonia species, but it also characterized the consequences of these genomic differences on gene expression. The genetic basis of learning and memory formation are highly conserved among distant animal species and insights from this thesis are likely applicable to other animal species and humans, as well.Altogether, these small parasitic wasps allow us to understand and value differences in memory formation.

The ability to learn and form memory has been demonstrated in various animal species, ranging from relatively simple invertebrates, such as snails and insects, to more complex vertebrate species, including birds and mammals. The opportunity to acquire new skills or to adapt behaviour through learning is an obvious benefit. However, memory formation is also costly: it can be maladaptive when unreliable associations are formed and the process of memory formation can be energetically costly. The balance between costs and benefits determines if learning and memory formation are beneficial to an animal or not. Variation in learning abilities and memory formation between species is thought to reflect species-specific differences in ecology. This thesis focused on variation in the number of trials required to form long-term memory (LTM). LTM is considered the most stable and durable type of memory, but also the most costly, because it requires protein synthesis. Many animal species require multiple learning experiences, which are spaced in time, to form LTM. This allows re-evaluation of information before an animal invests in costly LTM. There is, however, variation in the number of trials that animal species require to induce LTM formation. A number of insect species, including a number of parasitic wasp species, form LTM after only a single learning experience. Parasitic wasps can learn odours that guide them towards suitable hosts for their offspring, so-called oviposition learning. Substantial differences in LTM formation are observed among closely related species of parasitic wasps, which provides excellent opportunities for comparative studies. Both ecological and genetic factors involved in variation in LTM formation have been studied in this project. A multidisciplinary approach is essential to understand the evolution of variation in LTM formation, because the interaction between genes and environment shapes learning and memory formation. LTM formation was studied in closely related species of the genus Nasonia. These small parasitic wasps (~2 mm in length) lay their eggs in various species of fly pupae and differences in the ecology of the four known species of this genus have been described. A high-throughput method for olfactory conditioning was developed in which the wasps associated an odour, either chocolate or vanilla, with the reward of a host. A T-maze olfactometer was designed for high-throughput testing of memory retention. Using these methods, variation in memory retention was observed between three Nasonia species. Both N. vitripennis and N. longicornis form a long-lasting memory after a single conditioning trial, which lasts at least 5 days. Nasonia giraulti, on the other hand, lost its memory after 1 to 2 days after a single conditioning trial. Further studies focused on the difference between N. vitripennis and N. giraulti, which was most pronounced. By inhibiting LTM with transcription and translation inhibitors, it was confirmed that N. vitripennis forms this type of memory after a single conditioning trial. LTM is visible 4 days after conditioning in N. vitripennis. Nasonia giraulti does not form LTM after a single conditioning trial. Long-lasting memory is only formed after two trials, with a 4-hour interval between them. This difference in LTM formation makes N. vitripennis and N. giraulti excellent model species to study both ecological and genetic factors involved in this difference. Ecological factors such as the value of the reward and the reliability of the learned association have been shown to affect memory formation in a number of animal species. A recent study on oviposition learning in two parasitic wasp species demonstrated that LTM formation depends on the host species, i.e. the reward offered during conditioning. LTM was formed when a host with a higher quality was offered, but not when a host of lower quality was offered. The effect of host quality on memory retention of N. vitripennis and N. giraulti was tested. Either a large host, Calliphora vomitoria, a medium-sized host, Lucilia sericata, or a small host, Musca domestica, was offered during conditioning. These hosts were observed to differ significantly in their quality, i.e. in the number of parasitoid offspring that emerged and the size of the offspring. There was, however, no effect of host species on memory retention in either Nasonia species. These results suggest that host quality is not important for LTM formation in N. vitripennis and N. giraulti. This observation shows that ecological factors that are important for memory formation in one species may not be important for another species. The genetic basis of memory formation is highly conserved among distant animal phyla. A large number of genes involved in LTM formation have been identified in genetic model organisms, including fruit flies, honeybees, the California sea hare, mice and rats, and the zebra finch.Genetic factors responsible for natural variation in LTM formation between species are currently unknown, however. Two approaches were used to study genetic factors responsible for the difference in LTM formation between N. vitripennis and N. giraulti. The first approach took advantage of the unique possibility to interbreed Nasonia species. Hybrid offspring of N. vitripennis and N. giraulti did not form LTM after a single conditioning trial, similar to N. giraulti. The dominant LTM phenotype of N. giraulti was then backcrossed into the genetic background of N. vitripennis for up to 5 generations. Using a genotyping microarray analysis and subsequent confirmation experiments, we detected two genomic regions (quantitative trait loci – QTLs) that both reduce long-lasting memory, but not completely remove this memory. These results indicate that multiple QTLs regulate the difference in LTM formation between the two Nasonia species. Concluding, our approach has provided insights in the genomic basis of a naturally occurring difference in LTM formation between two species. Excellent opportunities for fine-scale QTL mapping are available for the genus Nasonia. This will allow identification of decisive regulatory mechanisms involved in LTM formation that are located in the two genomic regions detected in this study. The second approach took advantage of next-generation sequencing techniques that allow transcriptome-wide studies of gene expression levels. RNA from heads of N. vitripennis and N. giraulti was collected before conditioning and immediately, 4 hours, or 24 hours after conditioning. This RNA was sequenced strand-specifically using HiSeq technology, which allows detection of sense and antisense transcripts. Various genes, from a number of different signalling pathways known to be involved in LTM formation, were uniquely differentially expressed after conditioning in N. vitripennis. These genes are likely involved in the ongoing process of LTM formation in this species. A number of other genes with a known role in LTM formation,including genes involved in dopamine synthesis and in the Ras-MAPK and PI3K signalling pathways, were uniquely differentially expressed in N. giraulti. These genes may have a role in a LTM inhibitory mechanism in this species. Antisense transcripts were detected for a number of known memory genes, which may indicate a role inregulation of transcription, alternative splicing, or translation. This study is the first to compare gene expression patterns after conditioning between two species that differ in LTM formation. The results provide promising candidate genes for future studies in which the regulation of these genes, the function of specific splice variants, and spatial expression patterns in the brain should be studied to understand how these genes are involved in the regulation of LTM formation. Learning and memory formation have an important role in animal and human behaviour.Novel and valuable insights on both ecological and genetic factors responsible for variation in LTM formation have been revealed by the research presented in this thesis. Integrating ecological factors and genetic factors is essential, as genes are the level on which ecological factors can drive the evolution of variation in learning and memory formation. The genus Nasonia has offered excellent opportunities for ecological research as well as unique opportunities for studies on genomic and genetic factors, which were addressed by comparing closely related species that differ in memory formation. This thesis provides the basis for the identification of genomic differences responsible for the difference in memory formation between Nasonia species, but it also characterized the consequences of these genomic differences on gene expression. The genetic basis of learning and memory formation are highly conserved among distant animal species and insights from this thesis are likely applicable to other animal species and humans, as well.Altogether, these small parasitic wasps allow us to understand and value differences in memory formation.

Veel planten zijn van nature daglengtegevoelig. Ze gaan pas bloeien als de dag kort of juist lang genoeg is. Daarvoor hebben ze een klok nodig om de tijd te kunnen meten en een sensor om te kunnen zien of het licht is. De laatste jaren groeit het inzicht in de genetische aansturing sterk. Dat maakt de zaak ingewikkelder, maar geeft tevens aanknopingspunten voor mogelijke teeltmaatregelen in de toekomst.

De kans op obesitas is voor een aanzienlijk deel erfelijk bepaald. Welke specifieke genetische factoren daarbij betrokken zijn, is nog niet bekend. Daarom is het op dit moment niet gerechtvaardigd om genetische informatie te gebruiken om obesitas te voorkomen of te behandelen. Dit is de conclusie van een literatuurstudie naar genetische factoren voor obesitas die in opdracht van het ministerie van VWS is uitgevoerd door het Rijksinstituut voor Volksgezondheid en Milieu (RIVM). Het onderzoek heeft aangetoond dat ongeveer 40% van de variatie in het lichaamsgewicht tussen personen (uitgedrukt in de body mass index, BMI: gewicht / lengte-2) verklaard kan worden door genetische verschillen. Ook reageren mensen door hun erfelijk materiaal verschillend op veranderingen in energie-inname, -verbruik. Van een op de tien gevallen die op zeer jonge leeftijd extreem obees zijn, is een zeldzame genmutatie bekend. Van de minder extreme vormen van obesitas is hooguit 2,5% hiermee te verklaren. Veel onderzoek is gedaan naar variaties in genen die bij veel mensen voorkomen. Van vijf genvarianten die bekend zijn, is overtuigend bewezen dat ze van invloed zijn op de BMI en het risico op obesitas. Zij verklaren mogelijk 10% van de gevallen met overgewicht en 20% van de gevallen met obesitas. De genetische verschillen tussen personen zijn echter complexer dan voorheen werd gedacht. Daarom is naar verwachting de genetische bijdrage aan het ontstaan van obesitas groter dan momenteel bekend is. Dit rechtvaardigt lopend en toekomstig onderzoek op dit terrein.

At Swine Innovation Centre Sterksel the effects of genetic background, birth weight and feeding strategy during the weaning period on the performance and financial results of piglets and growing and finishing pigs were investigated. The results are described in this report.

Common sole (Solea solea) has a high potential for commercial aquaculture because of its consumer popularity and high market values in Europe. However, a major economic constraint for the culture of sole is its slow and variable growth. The aim of this thesis was to investigate: 1) the importance of (non-) feeding behaviour of sole in relation to variation in growth; 2) the effect of (social and physical) environmental factors on behaviour, growth and the relation between them; 3) the existence of GE interaction regarding growth. Feeding consistency, swimming activity in the tank, and boldness during (novel environment and light avoidance) behavioural tests explained variation in feed intake and thereby growth of individually housed sole. For communally housed sole, behavioural factors derived from individual behavioural tests and sex also explained variation in growth. The motivation to bury was negatively related to growth, whereas the motivation to explore a novel environment was positively related. Social interactions, both in quality (i.e., size hierarchies) and in quantity (i.e.,stocking density), influenced (non-) feeding behaviour and growth of sole. High stocking density in sole reared without substrate results in more fish-fish interactions, which increases swimming activity, FCR and variation in growth. These conditions seem to induce social stress in sole, which is alleviated when sand is provided. Environmental factors which differ between nature and farming conditions, such as food type, sand and variability of environmental conditions, influenced individual behavioural responses of sole to a novel environment test but did not induce variation in growth. Results suggest that consistent relationships between behaviour and growth develop when fish are reared in stable barren environments but not when fish experience more variable, enriched/natural environments. The role of environmental factors in the relationship between (non)-feeding behaviour and growth was supported by strong genotype by environment interaction for growth of sole reared in a semi-natural or an intensive aquaculture environment. In conclusion, the effect of (non-) feeding behaviour on growth should be taken into account to foster progress in the farming of sole. Environmental factors (i.e., substrate, stocking density) that influence behaviour and growth should be used to optimize culture systems. Future genetic selection strategies should focus more on behavioural characteristics to select sole which will be able to cope and grow best in the different rearing conditions present in commercial aquaculture.

At Swine Innovation Centre Sterksel the effects of genetic background, birth weight and feeding strategy during the weaning period on the performance and financial results of piglets and growing and finishing pigs were investigated. The results are described in this report. Op VIC Sterksel is onderzocht wat het effect is van de genetische aanleg van de eindbeer en van de zeug, het geboortegewicht van de biggen en de voerstrategie tijdens de opfokperiode op de technische en economische resultaten van biggen en vleesvarkens. De resultaten van het onderzoek zijn in dit rapport beschreven.

Common sole (Solea solea) has a high potential for commercial aquaculture because of its consumer popularity and high market values in Europe. However, a major economic constraint for the culture of sole is its slow and variable growth. The aim of this thesis was to investigate: 1) the importance of (non-) feeding behaviour of sole in relation to variation in growth; 2) the effect of (social and physical) environmental factors on behaviour, growth and the relation between them; 3) the existence of GE interaction regarding growth. Feeding consistency, swimming activity in the tank, and boldness during (novel environment and light avoidance) behavioural tests explained variation in feed intake and thereby growth of individually housed sole. For communally housed sole, behavioural factors derived from individual behavioural tests and sex also explained variation in growth. The motivation to bury was negatively related to growth, whereas the motivation to explore a novel environment was positively related. Social interactions, both in quality (i.e., size hierarchies) and in quantity (i.e.,stocking density), influenced (non-) feeding behaviour and growth of sole. High stocking density in sole reared without substrate results in more fish-fish interactions, which increases swimming activity, FCR and variation in growth. These conditions seem to induce social stress in sole, which is alleviated when sand is provided. Environmental factors which differ between nature and farming conditions, such as food type, sand and variability of environmental conditions, influenced individual behavioural responses of sole to a novel environment test but did not induce variation in growth. Results suggest that consistent relationships between behaviour and growth develop when fish are reared in stable barren environments but not when fish experience more variable, enriched/natural environments. The role of environmental factors in the relationship between (non)-feeding behaviour and growth was supported by strong genotype by environment interaction for growth of sole reared in a semi-natural or an intensive aquaculture environment. In conclusion, the effect of (non-) feeding behaviour on growth should be taken into account to foster progress in the farming of sole. Environmental factors (i.e., substrate, stocking density) that influence behaviour and growth should be used to optimize culture systems. Future genetic selection strategies should focus more on behavioural characteristics to select sole which will be able to cope and grow best in the different rearing conditions present in commercial aquaculture.

Background Coronary heart disease (CHD) continues to be a leading cause of morbidity and mortality among adults worldwide. Deregulated lipid metabolism (dyslipidemia) that manifests as hypercholesterolemia, hypertriglyceridemia, low high-density-lipoprotein (HDL) cholesterol levels or a combination of those, is an established risk factor for CHD among other established risk factors. Linoleic acid (LA, C18:2n-6) and alpha-linolenic acid (ALA, C18:3n-3) are polyunsaturated fatty acids (PUFAs) that cannot be synthesized de novo by human or animal cells, and therefore must be obtained from the diet. From these two PUFAs, two series of long-chain PUFAs are formed; the omega-6 series that are synthesized from LA, and the omega-3 series that are from ALA. Formation of these long-chain PUFAs involves a series of alternate desaturation and elongation processes. These PUFAs, especially, omega-3 PUFAs, have long been observed to reduce CHD risk. In contrast to the consistently observed cardiovascular protective effects of omega-3 PUFAs, accumulating evidence suggests a potential pro-atherogenic effects of omega-6 PUFAs, which is now still under debate. It has been estimated that genetic factors account for 26%-69% of inter-individual variation in CHD risk. These genetic factors are thought to influence CHD risk both directly and through effects on known CHD risk factors such as plasma lipid levels. The heritability of plasma lipid levels (total cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides (TG)) is estimated to be about 50% (ranging from 28%-78%). With the success of recent genome-wide association studies (GWAS), many genetic variants underlying intermediate risk factors of CHD (including plasma lipid levels) and CHD itself have been identified. Whether this new genetic information could be used to improve CHD risk prediction is still marginally explored, and for some variants, the underlying mechanisms for their mediated effects on CHD risk are still unknown. The aim of this research is to investigate common genetic determinants of plasma lipid levels (cholesterol and polyunsaturated fatty acid levels) using a pathway-driven approach, and to explore whether such common genetic variants could be used to improve CHD prediction using a population based genetic approach. An additional aim was to explore the underlying mechanisms of cardiovascular protective effects of PUFAs using a genomic approach. Methods In order to explore whether common genetic variants are involved in determining plasma cholesterol levels, we used data from 3575 men and women from the Doetinchem cohort, which was examined thrice over 11 years. They were genotyped on 384 single nucleotide polymorphisms (SNPs) across 251 genes in regulatory pathways that control fatty acid, glucose, cholesterol and bile salt homeostasis. In order to explore whether common genetic variants could be used to predict future CHD risk,we used the data from CAREMA cohort that involved 15,236 middle-aged subjects and was followed up for a median of 12.1 years. 179 SNPs associated with CHD or its risk factors in GWAS published up to May 2, 2011 were genotyped in the 2221 subcohort members and 742 incident CHD cases. In addition, fatty acids from plasma cholesteryl esters were quantified in 1323 subcohort members and 537 CHD cases. They were used to explore whether δ-5 and δ-6 desaturase activities were associated with CHD risk. In order to perform a comparative analysis of the effects of fenofibrate and fish oil at transcriptome and metabolome level, 34 mice were randomized by weight-matching into three groups (n = 10 in control group, and n = 12 in fenofibrate or fish oil intervention group), and fed a research diet supplemented with sunflower oil (containing 81.3% oleic acid, 7% energy intake) in control group, sunflower oil (7% energy intake) and fenofibrate (0.03% w/w) in fenofibrate group, and fish oil (Marinol C-38 fish oil: 23.1% EPA and 21.1% DHA, 7% energy intake) in fish oil group for 2 weeks. At the end of treatment, mice were fasted with drinking water available, and were subsequently sacrificed by cervical dislocation under isoflurane anesthesia. Blood was collected via orbital puncture. Livers were dissected, directly frozen in liquid nitrogen and stored at −80°C until further analysis. Microarray analysis was performed on individual mouse livers. The LC-MS method was used for measuring plasma lipids and non-esterified free fatty acids, and the GC-MS method was used for measuring a broad range of metabolites. Results In chapter 2, 3, and 4, common genetic variants in the genes along known cholesterol metabolic pathways, such as bile acid and bile metabolic pathways, the HDL cholesterol metabolic pathway, and the plasma total cholesterol metabolic pathway, are involved in determining plasma cholesterol levels. The modest effect associated with each individual variant, however, caused the amount of heritability explained by them in aggregate to be relatively small: 13 single nucleotide polymorphisms (SNPs) explained 4% of inter-individual variation in HDL cholesterol levels (Chapter 3), whereas 12 SNPs explained 6.9% of inter-individual variation in total cholesterol levels (Chapter 4). In chapter 5, we found that genetic variants in the FADS1 gene potentially interact with dietary PUFA intake to affect plasma cholesterol levels. A high intake of omega-3 PUFAs was associated with increased plasma non-HDL cholesterol levels, consistent with increased plasma LDL cholesterol levels observed in fish oil intervention studies. Increased LDL cholesterol levels could be due to hepatic downregulation of the LDL receptor gene (LDLR) in subjects with high omega-3 PUFA intakes. This is further confirmed by the findings described in Chapter 6 that the hepatic LDLR gene was significantly downregulated in fish oil treated mice. This study also confirmed PUFAs to be weak PPAR ligands. The increased plasma HDL cholesterol levels in the subjects with high PUFA intakes in Chapter 5 could be due to PPARs-mediated genes that are directly involved in HDL lipoprotein metabolism. All these may explain the changes in blood cholesterol levels upon PUFA intake observed in human studies. In Chapter 6, we found that not only downregulation in the hepatic lipogenic pathway but also upregulation in hepatic fatty acid oxidation pathways are involved in lowering plasma TG levels upon fish oil treatment. The striking parallel between fenofibrate and fish oil in hepatic downregulation of blood coagulation and fibrinolysis pathways suggest that hepatic activation of PPARα is potentially one of the mechanisms responsible for anticoagulation effects of fish oil treatment observed in humans. In Chapter 7, with confirmed effects of rs174547 in FADS1 on PUFA levels and δ-5 desaturase activities and also protective effects of DHA on CHD, we observed a reduced CHD risk of increased δ-5 desaturase activity. Increased δ-5 desaturase activity could contribute to the intracellular increase of EPA and especially arachidonic acid (C20:4n-6) levels. Despite the potential pro-coagulant and pro-inflammatory effects of increased exposures of arachidonic acid and its derived eicosanoid metabolites, there is no evidence of increased CHD risk with increased habitual arachidonic acid intake so far. Some of the oxygenated metabolites of arachidonic acid were found to have anti-inflammatory and pro-resolving actions. High dietary n-6 PUFA intakes or high plasma n-6 PUFA levels are associated with increased blood HDL cholesterol levels and reduced TG (or VLDL particle) levels. All these point to a potential cardiovascular protective effect of n-6 PUFAs. The fact that increased EPA and/or DHA levels associated with increased δ-5 desaturase activity protect against CHD is consistent with the current established cardiovascular protective effect of increased n-3 PUFA exposure, especially EPA and DHA. In Chapter 8, the current known common genetic variants associated with CHD risk factors (blood pressure, obesity, blood lipid levels, and type 2 diabetes) and CHD itself from published GWAS are examined to see whether they provide additional value in CHD risk prediction beyond established traditional CHD risk factors. We constructed several gene risk scores (GRS) for CHD that consisted of SNPs directly associated with CHD or intermediate CHD risk factors in GWAS, and tested their relationship to incident CHD and their potential to improve risk prediction. The weighted GRS based on 29 CHD SNPs predicted future CHD independently from established traditional risk factors. However, the GRS based on 153 SNPs associated with intermediate risk factors and the GRS based on the total 179 SNPs did not. None of them improved risk discrimination. Risk classification of CHD, measured by the net reclassification index, improved only when the GRS based on the 29 CHD SNPs was used. These results are generally consistent with the results from other recent studies that took a similar approach as ours. However, the final conclusions on GRS application could not be drawn at this early stage. With a great understanding of the genetic architecture of CHD in the future, more research should be done on this topic. Conclusions Our studies in this thesis demonstrated that common genetic variants along the known candidate cholesterol metabolic pathways are involved in determining the plasma cholesterol levels. PUFAs are not only weak PPARα ligands, but also inhibit SREBPs’ activities. All these could explain part of the cardiovascular protective effects (increased HDL cholesterol levels and reduced TG levels) of PUFAs, increased LDL cholesterol levels upon fish oil treatment in humans, and potentially reduced CHD risk of high δ-5 desaturase activities. At present, many questions remain about the feasibility of genetic risk prediction of CHD. Clinicians should continue to inquire about family history of CHD for risk prediction, because this represents a simple, cheap, and useful risk factor for CHD that likely represents the net integrated effects from hundreds of genetic risk variants.

Background Coronary heart disease (CHD) continues to be a leading cause of morbidity and mortality among adults worldwide. Deregulated lipid metabolism (dyslipidemia) that manifests as hypercholesterolemia, hypertriglyceridemia, low high-density-lipoprotein (HDL) cholesterol levels or a combination of those, is an established risk factor for CHD among other established risk factors. Linoleic acid (LA, C18:2n-6) and alpha-linolenic acid (ALA, C18:3n-3) are polyunsaturated fatty acids (PUFAs) that cannot be synthesized de novo by human or animal cells, and therefore must be obtained from the diet. From these two PUFAs, two series of long-chain PUFAs are formed; the omega-6 series that are synthesized from LA, and the omega-3 series that are from ALA. Formation of these long-chain PUFAs involves a series of alternate desaturation and elongation processes. These PUFAs, especially, omega-3 PUFAs, have long been observed to reduce CHD risk. In contrast to the consistently observed cardiovascular protective effects of omega-3 PUFAs, accumulating evidence suggests a potential pro-atherogenic effects of omega-6 PUFAs, which is now still under debate. It has been estimated that genetic factors account for 26%-69% of inter-individual variation in CHD risk. These genetic factors are thought to influence CHD risk both directly and through effects on known CHD risk factors such as plasma lipid levels. The heritability of plasma lipid levels (total cholesterol, LDL cholesterol, HDL cholesterol, and triglycerides (TG)) is estimated to be about 50% (ranging from 28%-78%). With the success of recent genome-wide association studies (GWAS), many genetic variants underlying intermediate risk factors of CHD (including plasma lipid levels) and CHD itself have been identified. Whether this new genetic information could be used to improve CHD risk prediction is still marginally explored, and for some variants, the underlying mechanisms for their mediated effects on CHD risk are still unknown. The aim of this research is to investigate common genetic determinants of plasma lipid levels (cholesterol and polyunsaturated fatty acid levels) using a pathway-driven approach, and to explore whether such common genetic variants could be used to improve CHD prediction using a population based genetic approach. An additional aim was to explore the underlying mechanisms of cardiovascular protective effects of PUFAs using a genomic approach. Methods In order to explore whether common genetic variants are involved in determining plasma cholesterol levels, we used data from 3575 men and women from the Doetinchem cohort, which was examined thrice over 11 years. They were genotyped on 384 single nucleotide polymorphisms (SNPs) across 251 genes in regulatory pathways that control fatty acid, glucose, cholesterol and bile salt homeostasis. In order to explore whether common genetic variants could be used to predict future CHD risk,we used the data from CAREMA cohort that involved 15,236 middle-aged subjects and was followed up for a median of 12.1 years. 179 SNPs associated with CHD or its risk factors in GWAS published up to May 2, 2011 were genotyped in the 2221 subcohort members and 742 incident CHD cases. In addition, fatty acids from plasma cholesteryl esters were quantified in 1323 subcohort members and 537 CHD cases. They were used to explore whether δ-5 and δ-6 desaturase activities were associated with CHD risk. In order to perform a comparative analysis of the effects of fenofibrate and fish oil at transcriptome and metabolome level, 34 mice were randomized by weight-matching into three groups (n = 10 in control group, and n = 12 in fenofibrate or fish oil intervention group), and fed a research diet supplemented with sunflower oil (containing 81.3% oleic acid, 7% energy intake) in control group, sunflower oil (7% energy intake) and fenofibrate (0.03% w/w) in fenofibrate group, and fish oil (Marinol C-38 fish oil: 23.1% EPA and 21.1% DHA, 7% energy intake) in fish oil group for 2 weeks. At the end of treatment, mice were fasted with drinking water available, and were subsequently sacrificed by cervical dislocation under isoflurane anesthesia. Blood was collected via orbital puncture. Livers were dissected, directly frozen in liquid nitrogen and stored at −80°C until further analysis. Microarray analysis was performed on individual mouse livers. The LC-MS method was used for measuring plasma lipids and non-esterified free fatty acids, and the GC-MS method was used for measuring a broad range of metabolites. Results In chapter 2, 3, and 4, common genetic variants in the genes along known cholesterol metabolic pathways, such as bile acid and bile metabolic pathways, the HDL cholesterol metabolic pathway, and the plasma total cholesterol metabolic pathway, are involved in determining plasma cholesterol levels. The modest effect associated with each individual variant, however, caused the amount of heritability explained by them in aggregate to be relatively small: 13 single nucleotide polymorphisms (SNPs) explained 4% of inter-individual variation in HDL cholesterol levels (Chapter 3), whereas 12 SNPs explained 6.9% of inter-individual variation in total cholesterol levels (Chapter 4). In chapter 5, we found that genetic variants in the FADS1 gene potentially interact with dietary PUFA intake to affect plasma cholesterol levels. A high intake of omega-3 PUFAs was associated with increased plasma non-HDL cholesterol levels, consistent with increased plasma LDL cholesterol levels observed in fish oil intervention studies. Increased LDL cholesterol levels could be due to hepatic downregulation of the LDL receptor gene (LDLR) in subjects with high omega-3 PUFA intakes. This is further confirmed by the findings described in Chapter 6 that the hepatic LDLR gene was significantly downregulated in fish oil treated mice. This study also confirmed PUFAs to be weak PPAR ligands. The increased plasma HDL cholesterol levels in the subjects with high PUFA intakes in Chapter 5 could be due to PPARs-mediated genes that are directly involved in HDL lipoprotein metabolism. All these may explain the changes in blood cholesterol levels upon PUFA intake observed in human studies. In Chapter 6, we found that not only downregulation in the hepatic lipogenic pathway but also upregulation in hepatic fatty acid oxidation pathways are involved in lowering plasma TG levels upon fish oil treatment. The striking parallel between fenofibrate and fish oil in hepatic downregulation of blood coagulation and fibrinolysis pathways suggest that hepatic activation of PPARα is potentially one of the mechanisms responsible for anticoagulation effects of fish oil treatment observed in humans. In Chapter 7, with confirmed effects of rs174547 in FADS1 on PUFA levels and δ-5 desaturase activities and also protective effects of DHA on CHD, we observed a reduced CHD risk of increased δ-5 desaturase activity. Increased δ-5 desaturase activity could contribute to the intracellular increase of EPA and especially arachidonic acid (C20:4n-6) levels. Despite the potential pro-coagulant and pro-inflammatory effects of increased exposures of arachidonic acid and its derived eicosanoid metabolites, there is no evidence of increased CHD risk with increased habitual arachidonic acid intake so far. Some of the oxygenated metabolites of arachidonic acid were found to have anti-inflammatory and pro-resolving actions. High dietary n-6 PUFA intakes or high plasma n-6 PUFA levels are associated with increased blood HDL cholesterol levels and reduced TG (or VLDL particle) levels. All these point to a potential cardiovascular protective effect of n-6 PUFAs. The fact that increased EPA and/or DHA levels associated with increased δ-5 desaturase activity protect against CHD is consistent with the current established cardiovascular protective effect of increased n-3 PUFA exposure, especially EPA and DHA. In Chapter 8, the current known common genetic variants associated with CHD risk factors (blood pressure, obesity, blood lipid levels, and type 2 diabetes) and CHD itself from published GWAS are examined to see whether they provide additional value in CHD risk prediction beyond established traditional CHD risk factors. We constructed several gene risk scores (GRS) for CHD that consisted of SNPs directly associated with CHD or intermediate CHD risk factors in GWAS, and tested their relationship to incident CHD and their potential to improve risk prediction. The weighted GRS based on 29 CHD SNPs predicted future CHD independently from established traditional risk factors. However, the GRS based on 153 SNPs associated with intermediate risk factors and the GRS based on the total 179 SNPs did not. None of them improved risk discrimination. Risk classification of CHD, measured by the net reclassification index, improved only when the GRS based on the 29 CHD SNPs was used. These results are generally consistent with the results from other recent studies that took a similar approach as ours. However, the final conclusions on GRS application could not be drawn at this early stage. With a great understanding of the genetic architecture of CHD in the future, more research should be done on this topic. Conclusions Our studies in this thesis demonstrated that common genetic variants along the known candidate cholesterol metabolic pathways are involved in determining the plasma cholesterol levels. PUFAs are not only weak PPARα ligands, but also inhibit SREBPs’ activities. All these could explain part of the cardiovascular protective effects (increased HDL cholesterol levels and reduced TG levels) of PUFAs, increased LDL cholesterol levels upon fish oil treatment in humans, and potentially reduced CHD risk of high δ-5 desaturase activities. At present, many questions remain about the feasibility of genetic risk prediction of CHD. Clinicians should continue to inquire about family history of CHD for risk prediction, because this represents a simple, cheap, and useful risk factor for CHD that likely represents the net integrated effects from hundreds of genetic risk variants.

Veel planten zijn van nature daglengtegevoelig. Ze gaan pas bloeien als de dag kort of juist lang genoeg is. Daarvoor hebben ze een klok nodig om de tijd te kunnen meten en een sensor om te kunnen zien of het licht is. De laatste jaren groeit het inzicht in de genetische aansturing sterk. Dat maakt de zaak ingewikkelder, maar geeft tevens aanknopingspunten voor mogelijke teeltmaatregelen in de toekomst.

WUR glastuinbouw heeft de alternatieven voor groeiremmers op een rijtje gezet. Een combinatie van maatregelen, die voor elk gewas anders ligt, kan het gebruik van groeiremmers terugdringen. Zij blijven echter wel nodig als correctiemiddel.

Woldwide, but especially in Europe, poultry husbandry will undergo significant changes due to the prohibition of both battery cage systems and beak-trimming. In laying hens, these changes will increase the risk of feather pecking. Feather pecking is defined as the non-aggressive pecking towards the plumage of other birds. It may result in feather damage and mortality due to cannibalism, which can be considered the ultimate phase of severe feather pecking. Feather pecking may therefore have negative consequences for bird welfare and the economic situation in poultry industry. To gain further insight in risk factors related to feather pecking, this thesis investigated the effects of genetic background and social environment on feather pecking and related behavioural characteristics in laying hens. In several experiments, behaviour, performance and physiology of cage-housed birds from pure-bred genetic lines was studied in different social environments at different ages. Results indicated that birds from different purebred lines show differences in feather damage due to severe feather pecking (an indicator for feather pecking) and in their response towards a novel object. This indicates that it is possible to select against high levels of both feaher pecking and fear related behaviour. The tendency to develop feather pecking was also related to the response towards a novel object, although this relation differed between birds from different backgrounds and from different ages. Other results showed that the response in the novel object test was also related to performance, which should be taken into account if such a test would to be used in a breeding program. Feather pecking and fear related behaviour were also affected by group mates (social environment): non-fearful birds became more fearful in presence of fearful birds. This effect could only be established at 18, but not at 5-6 weeks of age. At adult age, fearful birds showed more feather damage in presence of non-fearful birds, whereas the social environment during rearing had no effect on the occurrence of feather pecking. This indicates that fearful behaviour predisposes adult birds both to more easily develop and to be targeted by feather pecking. The changes in social environment were, however, not accompanied by physiological changes in brain serotonine or dopamine activity. These neurotransmission systems have been related to feather pecking. Results did indicate that the role of serotonin uptake does require further attention. According to the results from this thesis, laying hens should be kept in behavioural uniform groups to minimize the damage due to feather pecking. Additionally, reducing the expression of feather pecking could be achieved by breeding against expression of fearful behaviour, but possible correlated changes in performance should be accounted for. It remains to be investigated how the results with respect to social environment can be translated towards more extensive systems, such as floor-housing.

The aim of the project ‘The genetics of robustness in laying hens’ was to investigate nature and regulation of robustness in laying hens under sub-optimal conditions and the possibility to increase robustness by using animal breeding without loss of production. At the start of the project, a robust animal was defined as ‘an animal under a normal physical condition that has the potential to keep functioning and take short periods to recover under varying environmental conditions’. Next, parameters or traits were selected that could give an indication of robustness, and that could be implemented into a breeding goal for robustness. The experiments described in this thesis investigated parameters of interest for robustness in laying hens, where the influence of genetic background, environmental conditions, and early-life experiences was used as framework. The first experiment aimed at genetic differences in innate (natural) humoral immune components between 12 purebred layer Lines. Levels of innate and specific immune competence depended on genotype. Within layer lines, however, innate immune competence was related with survival. The second experiment aimed at influence of, or response to, environmental conditions, i.e., climatic stress (high temperature) and microbial challenge (lipopolysaccharide). Comparable response patterns to climatic stress and microbial challenge were found within lines, but lines differed in response levels towards these stressors. The third experiment aimed at improvement of robustness by early-life experiences to climatic stress and microbial challenge. The data did not reveal improvement of robustness by early-life experiences. Results from these studies indicate that robustness mainly depends on genetic background and environmental circumstances and to a minor extent on early-life experiences. The basic elements for robustness are survival, reproduction, and responsiveness towards environmental stressors, where a robust laying hen is a hen with a high survival rate, high production level, and low responsiveness towards environmental stressors. We have established a predictive value for the level of NAb binding to KLH for survival of the laying period of laying hens. Besides, NAb have a moderate heritability, giving opportunity for selection towards this trait. Performance parameters and innate immune parameters are most likely not related and selection on innate immune parameters will probably not be on the expense of hen-day egg production. Implementation of selection for NAb into a breeding goal might, therefore, improve robustness of laying hens.

The aim of the project ‘The genetics of robustness in laying hens’ was to investigate nature and regulation of robustness in laying hens under sub-optimal conditions and the possibility to increase robustness by using animal breeding without loss of production. At the start of the project, a robust animal was defined as ‘an animal under a normal physical condition that has the potential to keep functioning and take short periods to recover under varying environmental conditions’. Next, parameters or traits were selected that could give an indication of robustness, and that could be implemented into a breeding goal for robustness. The experiments described in this thesis investigated parameters of interest for robustness in laying hens, where the influence of genetic background, environmental conditions, and early-life experiences was used as framework. The first experiment aimed at genetic differences in innate (natural) humoral immune components between 12 purebred layer Lines. Levels of innate and specific immune competence depended on genotype. Within layer lines, however, innate immune competence was related with survival. The second experiment aimed at influence of, or response to, environmental conditions, i.e., climatic stress (high temperature) and microbial challenge (lipopolysaccharide). Comparable response patterns to climatic stress and microbial challenge were found within lines, but lines differed in response levels towards these stressors. The third experiment aimed at improvement of robustness by early-life experiences to climatic stress and microbial challenge. The data did not reveal improvement of robustness by early-life experiences. Results from these studies indicate that robustness mainly depends on genetic background and environmental circumstances and to a minor extent on early-life experiences. The basic elements for robustness are survival, reproduction, and responsiveness towards environmental stressors, where a robust laying hen is a hen with a high survival rate, high production level, and low responsiveness towards environmental stressors. We have established a predictive value for the level of NAb binding to KLH for survival of the laying period of laying hens. Besides, NAb have a moderate heritability, giving opportunity for selection towards this trait. Performance parameters and innate immune parameters are most likely not related and selection on innate immune parameters will probably not be on the expense of hen-day egg production. Implementation of selection for NAb into a breeding goal might, therefore, improve robustness of laying hens.

Keywords:Pathogen-specific clinical mastitis, Somatic cell count patterns, Genetic selection programs, Udder health managementResearch described in the thesis provideinsight in the use of patterns of peaks in somatic cell count (scc) in genetic selection and mastitis control programs. Patterns of peaks insccduring a lactation were defined based onsccrecorded on consecutive test-days, and are based on biological understanding of pathogens and of the immune system of the cow. Results showed that selecting against occurrence ofsccpatterns was more effective to decrease the natural susceptibility to mastitis-causing pathogens, than selection for lower lactation-averagescc. Patterns of peaks insccare also proven to be useful as basic tools for health management advice, as they can distinguish between casesofmastitis associated with either environmental or contagious pathogens, whereas the currently used primary traits were indicative for contagious, but not for environmental mastitis.&nbsp

This thesis describes the genetic analysis of feather pecking behavior in laying hens. Feather pecking (FP) is a major welfare problem in laying hens.In the European Union, legislation concerning animal housing is becomingmore strict, because of increasing concern for animal welfare. Conventional cage systems will be banned in 2012 and there is a clear movement towards alternative housing systems such as modified cages and large group housing systems. Unfortunately, in large group housing systems FP is difficult to control. To study the genetics of FP behavior and open-field behavior, an F 2 population of 630 hens coming from two commercial laying lines was generated. All F 2 animals were tested for their FP behavior in a social FP test as well as for their behavior in an open-field test at both young and adult age. The F 0 , F 1 and F 2 generations were genotyped with 180microsatellitemarkers. Gentle FP was heritable at both young and adult age whereas severe FP was not. Heritability of open-field behavior at young age was higher than at adult age.A QTL analysis, using a half-sib model and a line-cross model was performed for different types of pecking behavior such as gentle FP, severe FP, aggressive pecking and toe pecking, as well as for receiving pecks and open-field behavior. For gentle FP suggestive QTL were identified at young age at GGA10 and at adult age at GGA2. In addition, for severe FP at adult age, a significant QTL was detected at GGA2. Parallel to this, using a comparative mapping approach, theglucocorticoidreceptor gene was mapped to GGA13. This chromosome was not detected in the QTL study which makes this gene less likely to be a candidate gene for FP behavior.For receiving gentle FP, a significant QTL was detected on GGA1 at young age and at adult age a suggestive QTL was detected on GGA5. For open-field behavior atyoungage a significant QTL was detected on GGA4 and at adult age a significant QTL was detected on GGA4 as well. The QTL are more than 100cM apart, therefore, it is not likely that these QTL share the same underlying genes. Interestingly, these results indicate that pecking behavior as well as open-field behavior at young age is regulated by different genes than at adult age. The QTL for severe FP and gentle FP offer the possibility to identify genetic markers for FP behavior, which may be used in genetic improvement programs.

Plasminogen activator inhibitor-type 1 (PAI-1) is the main inhibitor of fibrinolysis and a potential risk factor for cardiovascular disease. In addition to its regulating role in fibrinolysis, PAI-1 may have detrimental effects in the cardiovascular system also through other processes, e.g. inflammation. Although PAI-1 is generally elevated in the presence of cardiovascular disease, it is not yet clear whether it is a causal risk factor. A polymorphism in the promoter region of the PAI-1 gene, the 4G/5G-polymorphism, affects the transcription of the PAI-1 gene, yielding higher blood levels of PAI-1 for the 4G-allele. In case of a causal role of PAI-1 in cardiovascular disease, an increased cardiovascular risk would be expected for the 4G-allele.The general objective of the studies in this thesis was to examine whether PAI-1 and the 4G/5G-polymorphism are associated with cardiovascular risk. Both associations with markers of atherosclerosis and clinical end points were studied. Because of the acute-phase properties of PAI-1, we additionally examined the role of C-reactive protein (CRP), a sensitive marker of inflammation.In the Arnhem Elderly Study, comprising 641 subjects aged 65-84 years, we observed a strong daytime fluctuation in PAI-1 activity with peak levels in the early morning. The diurnal pattern was clearly present for the 4G/4G and 4G/5G-genotypes, but not for the 5G/5G-genotype. These findings raised the hypothesis that 5G-homozygotic persons may be relatively protected from the early morning peak incidence in cardiovascular events.In a population of 208 smoking men the 4G/5G-polymorphism was not associated with two non-invasive markers of atherosclerosis, i.e. common carotid intima-media thickness and the ankle-brachial index. Neither did we observe consistent associations between the 4G/5G-polymorphism and coronary stenosis after pooling three similar case-control studies (n = 776) with angiographically determined coronary atherosclerosis.In the Arnhem Elderly Study we investigated whether PAI-1 and the 4G/5G-polymorphism were associated with mortality and incidence of cardiovascular events during 7.8 years of follow-up. We observed that the 4G/4G-genotype was protective against incident stroke and transient ischemic attack (relative risks of 0.4 and 0.3 respectively), which is in contrast to the generally accepted hypothesis that the 4G-allele would increase cardiovascular risk. For PAI-1 levels, an increased risk of stroke and ischemic attacks was observed in the highest tertile.In the Arnhem Elderly Study, PAI-1 and CRP were inter-related, but only in lean elderly. The associations between CRP and other components of the metabolic syndrome were also predominantly present in lean elderly. In elderly men, CRP was predictive for future cardiovascular events, but not in women.Overall, PAI-1 and the 4G/5G-polymorphism do not appear to play a major role in atherosclerosis, but the effects on thrombosis are still controversial. Causality of PAI-1 was not supported by an increased risk of cardiovascular events for the 4G/4G-genotype. In contrast, subjects with the 4G/4G-genotype were relatively protected against stroke.

Keywords:Pathogen-specific clinical mastitis, Somatic cell count patterns, Genetic selection programs, Udder health managementResearch described in the thesis provideinsight in the use of patterns of peaks in somatic cell count (scc) in genetic selection and mastitis control programs. Patterns of peaks insccduring a lactation were defined based onsccrecorded on consecutive test-days, and are based on biological understanding of pathogens and of the immune system of the cow. Results showed that selecting against occurrence ofsccpatterns was more effective to decrease the natural susceptibility to mastitis-causing pathogens, than selection for lower lactation-averagescc. Patterns of peaks insccare also proven to be useful as basic tools for health management advice, as they can distinguish between casesofmastitis associated with either environmental or contagious pathogens, whereas the currently used primary traits were indicative for contagious, but not for environmental mastitis.&nbsp