Search

Lily (Lilium L.) is a perennial bulbous ornamental, belonging to subclass Monocotyledonae and family Liliaceae. Lily, according to statistics of Dutch auctions, is the fifth most important cut flower and the second in flower bulbs based on acreage. This species has been extensively used for cytogenetic studies, but molecular genetic studies are limited. The heterogenic nature and the very complex and huge genome (36 Gb) of lily might be the reason for this. To improve the efficiency of breeding and selection in this species, and set up the basis for genetic studies in Lilium, genomic resources are needed. Next generation sequencing (NGS) technology (454 pyro-sequencing) was used to sequence the transcriptomes (RNA-seq) of four lily cultivars: ‘Connecticut King’, ‘White Fox’, ‘Star Gazer’, and Trumpet that belong to the four most important hybrid groups: Asiatic, Longiflorum, Oriental, and Trumpet respectively. Successfully, 52,172 unigenes with an average length of 555 bp were developed and used for a wide range of genetic and genomic studies: SNP marker identification for genetic mapping, gene annotation, and comparative genomic studies. Combining NGS with SNP genotyping techniques to accelerate genetic studies is of considerable interest in different species. In this study, thousands of SNPs out of the 52,172 lily unigenes were identified. Genotyping technique KASPar (KBiosciences competitive Allele Specific PCR) was used to genotype two lily mapping populations: ‘LA (L. longiflorum ‘White Fox’ x Asiatic hybrid ‘Connecticut King’) and AA (‘Connecticut King’ x ‘Orlito’) using 225 SNP markers selected from ‘Connecticut King’ unigenes. Genotyping success rate was 75.5% (170 SNP markers worked), polymorphic SNP rate was 45% (102 SNP markers), and mapped SNP marker rate was 42% (94 SNP mapped) in LA population and 38% (85 SNP mapped) in AA population. Thus, we validated a subset of the putative SNP makers and showed the usability of this type of markers to improve genetic maps for complex genomes like that of lily. The SNP markers together with the available AFLP (amplified fragment length polymorphisms), DArT (diversity arrays technology), and NBS (nucleotide binding site) markers were used to build reference genetic maps for these two lily populations. These maps represent the first reasonably saturated maps that cover 89% of the lily genome with an average marker density of one marker per 4 cM. The availability of more SNP markers for genotyping, opens the door for further enriching these genetic maps and thus improve the marker density. The genetic maps were used to map and understand the genetic of several horticultural traits in Lilium. Fusarium oxysporum and lily mottle virus (LMoV) are considered as very serious diseases in Lilium and as such present important targets for breeding. Six putative QTLs (quantitive trait loci) were identified for Fusarium resistance in AA population, from which QTL1 was the strongest (explains ~25 % of phenotypic variation). In LA population, QTL1 was also confirmed. Thus, QTL1 is a strong and reliable QTL in both populations and it can be used to develop markers for most of the Fusarium resistance for molecular assisted breeding (MAB) applications. The LMoV was mapped as a marker on the AA genetic maps, however, no close markers to this trait (i.e. distance of the closest marker to LMoV was 9 cM) were identified yet. Several ornamental traits: lily flower color ‘carotene’ (LFCc), flower spots (lfs), stem color (LSC), antherless phenotype (lal), and flower direction (up-side facing, lfd) were phenotyped and mapped. Some of these traits showed to be recessive traits (spots, antherless, and flower direction) and controlled by a single gene. Developing markers for recessive traits is valuable since such markers allow the identification of suitable breeding parents so the presence of the recessive trait can be either enhanced or repressed. A more complex trait is flower longevity because it is a function of: the number of buds per inflorescence, the expansion and opening of the buds, the life span of the individual flowers, and also the life-span of the leaves. Moreover, senescence in Lilium is ethylene-insensitive and the regulator(s) of its vase life is not known yet. Our study showed that vase life of individual lily flowers increased significantly by the exogenous application of sugars. Abscisic acid (ABA) level, furthermore, increased dramatically in lily flowers at senescence compared with anthesis. This indicates that ABA might be the main regulator of vase life in lily. However, more experiments should be conducted to prove this conclusion. The genomic resources developed for lily together with genomic resources developed for Tulipa L., in the same way, offered a valuable source of information to conduct comparative genomic studies within and between these two genera. We initiated the first step towards linking molecular genetic maps of Lilium and Tulipa using transcriptome sequences generated by 454 pyro-sequencing. Orthologous genes between lily and tulip were identified (10,913 unigenes) based on sequence data of four lily cultivars and five tulip cultivars. Next, common SNP and EST-SSR markers between the parents of lily mapping populations (AA and LA population) and the parents of tulip mapping population (‘Kees Nelis’ (T. gesneriana) x ‘Cantata’ (T. fosteriana)) based on these orthologous sequences were generated. A total of 229 common SNP and 140 common EST-SSR markers were identified. Genotyping and mapping these markers in the populations of both genera will link the genetic maps of Lilium and Tulipa and thus allow insight into the preservation of gene order, structure, and ‘putative’ functional homology in addition to evolutionary processes. Also, these genomic resources can be used to increase the resolution of, and support for, phylogenetic trees. We selected a set of orthologous genes of Lilium (19 genes, 11,766 bp containing 433 polymorphic sites), of Tulipa (20 genes, 10,347 bp containing 216 polymorphic sites), and of the orthologous genes between the two genera (7 genes, 5,790 bp containing 587 polymorphic sites). These sets are uniquely present in the sequences and informative in estimating the genetic divergence of the two genera, thus they can be used to genotypes more species per genera to build genera and maybe family trees later on. The nucleotide polymorphism rate of Lilium was twice as high as that of Tulipa, on average one substitution per 26 bp for Lilium compared with one substitution per 48 bp for Tulipa. NGS provide a valuable source for large numbers of phylogenetic informative substitutions that might revolutionize the phylogenetic, population genetic, and biodiversity studies. However, the use of bi-allelic information from multiple loci in phylogenetic studies is still challenging and it needs to be studied further. Moreover, having such high numbers of sequence data, allows us to test some evolutionary hypotheses such as positive selection: selection during domestication/breeding processes might be imprinted in the species genome, which can be examined based on omega (dn/ds) values. The higher the omega value the stronger the indication of positive selection. Positive selection was recorded in Lilium and Tulipa genomes when this small subset of gene contigs (46) of the two genera was tested. Our hypothesis could not be confirmed, however to draw final conclusions on this matter, omega values for many more genes of the two genera have to be measured. Finally, a wealth of putative molecular markers (SNPs and SSRs) has become available that can have direct applications for breeding in these genera. SNP markers are important since they are user friendly, efficient, transferable, and co-dominant markers. Applying high throughput genotyping technology to genotype the two lily populations improved the coverage of the two genetic maps. Also, genotyping the same SNP markers in the two populations facilitated the comparisons between the linkage groups of the two populations and will allow the construction of a consensus map. Consequently, exchange of genetic knowledge (mainly QTLs) between the two populations will be easier. The thousands of SNPs identified in the genome of the four lily cultivars opens the door for combining the current linkage mapping studies with association studies which will have a direct impact on improving the resolution of mapping and on MAB applications in Lilium.

Lily (Lilium L.) is a perennial bulbous ornamental, belonging to subclass Monocotyledonae and family Liliaceae. Lily, according to statistics of Dutch auctions, is the fifth most important cut flower and the second in flower bulbs based on acreage. This species has been extensively used for cytogenetic studies, but molecular genetic studies are limited. The heterogenic nature and the very complex and huge genome (36 Gb) of lily might be the reason for this. To improve the efficiency of breeding and selection in this species, and set up the basis for genetic studies in Lilium, genomic resources are needed. Next generation sequencing (NGS) technology (454 pyro-sequencing) was used to sequence the transcriptomes (RNA-seq) of four lily cultivars: ‘Connecticut King’, ‘White Fox’, ‘Star Gazer’, and Trumpet that belong to the four most important hybrid groups: Asiatic, Longiflorum, Oriental, and Trumpet respectively. Successfully, 52,172 unigenes with an average length of 555 bp were developed and used for a wide range of genetic and genomic studies: SNP marker identification for genetic mapping, gene annotation, and comparative genomic studies. Combining NGS with SNP genotyping techniques to accelerate genetic studies is of considerable interest in different species. In this study, thousands of SNPs out of the 52,172 lily unigenes were identified. Genotyping technique KASPar (KBiosciences competitive Allele Specific PCR) was used to genotype two lily mapping populations: ‘LA (L. longiflorum ‘White Fox’ x Asiatic hybrid ‘Connecticut King’) and AA (‘Connecticut King’ x ‘Orlito’) using 225 SNP markers selected from ‘Connecticut King’ unigenes. Genotyping success rate was 75.5% (170 SNP markers worked), polymorphic SNP rate was 45% (102 SNP markers), and mapped SNP marker rate was 42% (94 SNP mapped) in LA population and 38% (85 SNP mapped) in AA population. Thus, we validated a subset of the putative SNP makers and showed the usability of this type of markers to improve genetic maps for complex genomes like that of lily. The SNP markers together with the available AFLP (amplified fragment length polymorphisms), DArT (diversity arrays technology), and NBS (nucleotide binding site) markers were used to build reference genetic maps for these two lily populations. These maps represent the first reasonably saturated maps that cover 89% of the lily genome with an average marker density of one marker per 4 cM. The availability of more SNP markers for genotyping, opens the door for further enriching these genetic maps and thus improve the marker density. The genetic maps were used to map and understand the genetic of several horticultural traits in Lilium. Fusarium oxysporum and lily mottle virus (LMoV) are considered as very serious diseases in Lilium and as such present important targets for breeding. Six putative QTLs (quantitive trait loci) were identified for Fusarium resistance in AA population, from which QTL1 was the strongest (explains ~25 % of phenotypic variation). In LA population, QTL1 was also confirmed. Thus, QTL1 is a strong and reliable QTL in both populations and it can be used to develop markers for most of the Fusarium resistance for molecular assisted breeding (MAB) applications. The LMoV was mapped as a marker on the AA genetic maps, however, no close markers to this trait (i.e. distance of the closest marker to LMoV was 9 cM) were identified yet. Several ornamental traits: lily flower color ‘carotene’ (LFCc), flower spots (lfs), stem color (LSC), antherless phenotype (lal), and flower direction (up-side facing, lfd) were phenotyped and mapped. Some of these traits showed to be recessive traits (spots, antherless, and flower direction) and controlled by a single gene. Developing markers for recessive traits is valuable since such markers allow the identification of suitable breeding parents so the presence of the recessive trait can be either enhanced or repressed. A more complex trait is flower longevity because it is a function of: the number of buds per inflorescence, the expansion and opening of the buds, the life span of the individual flowers, and also the life-span of the leaves. Moreover, senescence in Lilium is ethylene-insensitive and the regulator(s) of its vase life is not known yet. Our study showed that vase life of individual lily flowers increased significantly by the exogenous application of sugars. Abscisic acid (ABA) level, furthermore, increased dramatically in lily flowers at senescence compared with anthesis. This indicates that ABA might be the main regulator of vase life in lily. However, more experiments should be conducted to prove this conclusion. The genomic resources developed for lily together with genomic resources developed for Tulipa L., in the same way, offered a valuable source of information to conduct comparative genomic studies within and between these two genera. We initiated the first step towards linking molecular genetic maps of Lilium and Tulipa using transcriptome sequences generated by 454 pyro-sequencing. Orthologous genes between lily and tulip were identified (10,913 unigenes) based on sequence data of four lily cultivars and five tulip cultivars. Next, common SNP and EST-SSR markers between the parents of lily mapping populations (AA and LA population) and the parents of tulip mapping population (‘Kees Nelis’ (T. gesneriana) x ‘Cantata’ (T. fosteriana)) based on these orthologous sequences were generated. A total of 229 common SNP and 140 common EST-SSR markers were identified. Genotyping and mapping these markers in the populations of both genera will link the genetic maps of Lilium and Tulipa and thus allow insight into the preservation of gene order, structure, and ‘putative’ functional homology in addition to evolutionary processes. Also, these genomic resources can be used to increase the resolution of, and support for, phylogenetic trees. We selected a set of orthologous genes of Lilium (19 genes, 11,766 bp containing 433 polymorphic sites), of Tulipa (20 genes, 10,347 bp containing 216 polymorphic sites), and of the orthologous genes between the two genera (7 genes, 5,790 bp containing 587 polymorphic sites). These sets are uniquely present in the sequences and informative in estimating the genetic divergence of the two genera, thus they can be used to genotypes more species per genera to build genera and maybe family trees later on. The nucleotide polymorphism rate of Lilium was twice as high as that of Tulipa, on average one substitution per 26 bp for Lilium compared with one substitution per 48 bp for Tulipa. NGS provide a valuable source for large numbers of phylogenetic informative substitutions that might revolutionize the phylogenetic, population genetic, and biodiversity studies. However, the use of bi-allelic information from multiple loci in phylogenetic studies is still challenging and it needs to be studied further. Moreover, having such high numbers of sequence data, allows us to test some evolutionary hypotheses such as positive selection: selection during domestication/breeding processes might be imprinted in the species genome, which can be examined based on omega (dn/ds) values. The higher the omega value the stronger the indication of positive selection. Positive selection was recorded in Lilium and Tulipa genomes when this small subset of gene contigs (46) of the two genera was tested. Our hypothesis could not be confirmed, however to draw final conclusions on this matter, omega values for many more genes of the two genera have to be measured. Finally, a wealth of putative molecular markers (SNPs and SSRs) has become available that can have direct applications for breeding in these genera. SNP markers are important since they are user friendly, efficient, transferable, and co-dominant markers. Applying high throughput genotyping technology to genotype the two lily populations improved the coverage of the two genetic maps. Also, genotyping the same SNP markers in the two populations facilitated the comparisons between the linkage groups of the two populations and will allow the construction of a consensus map. Consequently, exchange of genetic knowledge (mainly QTLs) between the two populations will be easier. The thousands of SNPs identified in the genome of the four lily cultivars opens the door for combining the current linkage mapping studies with association studies which will have a direct impact on improving the resolution of mapping and on MAB applications in Lilium.

In this thesis natural and induced DNA sequence diversity in potato (Solanum tuberosum) for use in marker-trait analysis and potato breeding is assessed. The study addresses the challenges of reliable, high-throughput identification and genotyping of sequence variants in existing tetraploid potato cultivar panels using traditional Sanger sequencing and next-generation massively parallel sequencing (MPS), and the application of this knowledge in the form of genetic markers. Furthermore, it explores the efficiency of ethyl methanesulphonate (EMS) mutagenesis combined with high resolution melting (HRM) DNA screening to induce and discover novel sequence variants in potato genotypes. Discovery and genotyping of sequence diversity in outcrossing autotetraploid species like potato is complex. In autotetraploid species, genotyping implies the quantitative identification of five alternative allele copy number states. In Chapter 1, several methodologies to identify and genotype DNA sequence variants, and the application of these sequence variants is discussed. This chapter provides an introduction to genotyping-by-sequencing (GBS) and the determination of allele copy number. In Chapter 2 the sequence diversity in three genes of the carotenoid pathway is assessed in diploid and tetraploid potato genotypes using direct Sanger sequencing. To investigate the genetics and molecular biology of orange and yellow flesh colour in potato, association analysis between SNP haplotypes and flesh colour phenotypes was performed, and the inheritance and gene expression of associated alleles was studied. We observed among eleven beta-carotene hydroxylase 2 (CHY2) alleles one dominant allele with a major effect, changing white into yellow flesh colour. In contrast, none of the lycopene epsilon cyclase (LCYe) alleles seemed to have a large effect on flesh colour. Analysis of zeaxanthin epoxidase (ZEP) alleles showed that a recessive allele with a non-LTR retrotransposon sequence in intron 1 reduced the expression level of the ZEP gene and caused accumulation of zeaxanthin. Genotypes combining presence of the dominant CHY2 allele with homozygosity for the recessive ZEP allele produced orange-fleshed tubers that accumulate large amounts of zeaxanthin. Sanger amplicon sequencing was applied in Chapter 3 to evaluate the sequence diversity in α-Glucan Water Dikinase (StGWD), a candidate gene underlying a QTL involved in potato starch phosphate content. Sanger sequences of two StGWD amplicons from a global collection of 398 commercial cultivars and progenitor lines were used to identify 16 unique haplotypes. By assigning tag SNPs to these haplotypes and by determining the allele copy number of identified sequence variants, we inferred the four-allele genetic composition for almost all cultivars assayed at this locus. This allowed genetic diversity parameters like the average number of different alleles present in a single cultivar (Ai=3.1) and the average intra-individual heterozygosity (Ho=0.765) to be estimated for this locus. Pedigree analysis confirmed that the identified haplotypes are identical by descent (IBD) and offered insight in the breeding history of elite potato germplasm. Haplotype association analysis led to the identification of two StGWD alleles causing altered starch phosphate content, which was further verified in diploid and tetraploid mapping populations containing the relevant alleles. One of these alleles (Allel H) increases the fraction of starch that is phosphorylated, while the other one (Allele A) decreases it. To scale up the discovery and genotyping of sequence variants, and to make it more whole-genome oriented, Chapter 4 reports on massively parallel sequencing (MPS) of approximately 800 genes scattered over the potato genome and resequenced in 83 tetraploid potato cultivars and a monoploid reference accession. We show that by combining MPS with genome complexity reduction and indexed sequencing, sufficient read depth for GBS can be achieved for reliable discovery and genotyping of sequence variants in individual tetraploid potato genotypes. With a custom designed, SureSelect enrichment library, 1.44 Mb of DNA sequence was targeted. The genes targeted were mainly single-copy genes, selected based on putative gene functions in both primary and secondary metabolic pathways, potato quality traits and biotic and abiotic stresses, and included a large set of conserved orthologous sequence genes (COSII) useful for genetic anchoring and phylogenetic studies. The indexed and enriched DNA libraries were sequenced on a Illumina HiSeq. After filtering and processing the raw sequence data, 12.4 Gb of high-quality sequence data was mapped to the potato genome, covering 2.1 Mb of the genome sequence with a median average read depth of 63× per cultivar. We detected over 129,000 sequence variants in these data and determined allele copy number of the variants in individual potato samples. The accuracy of the sequence-based allele copy number estimates was verified by a low-density SNP genotyping assay. This showed that for reliable genotyping a read count-based genotype quality score is best applied and a read depth of 80× is recommended for determining allele copy number in autotetraploid potato. Average nucleotide diversity (π=10.7×10-3 genome-wide, ≈1 variant/93 bp between two random alleles) varied along the twelve potato chromosomes, and individual genes under selection were identified. As an example for application of GBS for genome-wide association analysis (GWAS), the identified sequence variants and genotype data were tested in a marker-trait association analysis with plant maturity and tuber flesh colour. This led to the identification of alleles accounting for significant phenotypic variation in these traits. In Chapter 5 we applied the chemical mutagen EMS to diploid potato by two different treatments, a pollen and a seed treatment. We screened the resulting populations for novel mutations using HRM analysis. A pollen treatment with EMS dissolved in a sucrose solution was found to induce mutations only at a low frequency (only one mutation discovered after screening >2.7 Mb of sequence). In planta selection of the most vital mutagenized pollen seems to have lowered the mutation density to a frequency that is not suitable for reverse genetics studies. Treatment of potato seeds with EMS on the other hand provided a high density of novel mutations (1 mutation/65 kb), discovered in the M1 generation. In contrast to most EMS mutagenesis studies, we directly screened the M1 generation of the seed-treated population. In six candidate genes involved in potato starch and frying quality traits, 65 novel sequence variants were discovered. In all six genes, missense mutations that are predicted to damage protein function were discovered, and for four genes five premature stop codon mutations were identified. We attempted to stabilize and transfer 27 putatively interesting mutations to the M2 and M3 generation for further evaluation. Genetically stable M2 and M3 plants have been generated for 10 (37%) of these mutations. The estimated density of M1 mutations that are transferable to the M2 generation (one “accessible” mutation/118-176 kb) is higher than the mutation densities obtained in most other plant species, for which the M2 generation has been screened. The results of this chapter thus demonstrate that screening the M1 generation offers a good alternative to the commonly applied M2 screening for the rapid generation of novel genetic variation at a high density, without too much complication in recovering mutations in the M2 generation. In the concluding Chapter 6, results of preceding chapters are evaluated, and the prospects of the findings for potato research and breeding are discussed.

In this thesis we describe different approaches aiding in the utilization of the exponentially growing amount of information available in the life sciences. Briefly, we address two issues in molecular biology, on sequence analysis, and on text mining. The former issue addresses the problem how to determine remote sequence homology especially when the sequence similarity is very low. For this a visualisation tool is introduced that combines sequence alignment, domain prediction and phylogeny. The second topic on text mining centres on the question how to unambiguously formulate queries for efficient information retrieval. It tackles the problem of gene nomenclature — one in two gene symbols being ambiguous - by introducing a new text-clustering- and taxonomy-based disambiguation methodology.

In this thesis we describe different approaches aiding in the utilization of the exponentially growing amount of information available in the life sciences. Briefly, we address two issues in molecular biology, on sequence analysis, and on text mining. The former issue addresses the problem how to determine remote sequence homology especially when the sequence similarity is very low. For this a visualisation tool is introduced that combines sequence alignment, domain prediction and phylogeny. The second topic on text mining centres on the question how to unambiguously formulate queries for efficient information retrieval. It tackles the problem of gene nomenclature — one in two gene symbols being ambiguous - by introducing a new text-clustering- and taxonomy-based disambiguation methodology.

The thesis focuses on two bioinformatics research topics: the development of tools for an efficient and reliable identification of single nucleotides polymorphisms (SNPs) and polymorphic simple sequence repeats (SSRs) from expressed sequence tags (ESTs) (Chapter 2, 3 and 4), and the subsequent implementation of these tools in a pipeline for narrowing down QTL intervals to facilitate the identification of candidate genes for the QTL (Chapter 5). Chapter 1 provides an introduction to molecular markers, SNP and SSR markers, illustrated by a number of applications of molecular markers, as well as existing programs for their detection. After analysis of existing problems and programs for the detection SNPs, a new algorithm is described to reliably identify SNPs and indels in EST data from diploid and polyploid species (Chapter 2). The algorithm is implemented in a program called QualitySNP. This program uses three filters to identify reliable SNPs: filter one screens for potential SNPs by requiring at least two sequences per represented allele; filter two uses a haplotype-based strategy to filter out clusters with paralogs and false SNPs caused by sequence errors; finally, filter three calculates a confidence score for every putative SNP according to the number of occurrences of each allele in high and low quality regions. For the detection of non-synonymous SNPs (nsSNP), synonymous SNPs and SNPs in UTRs, a program was developed as well. Furthermore, these programs were implemented in a pipeline that includes the identification of SNPs and nsSNPs, as well as a storage and retrieval system. Using this pipeline, large numbers of SNPs could be identified in potato ESTs. QualitySNP is available for running on LINUX and UNIX systems; the program, user manual and examples are available at http://www.bioinformatics.nl/tools/snpweb/ . Chapter 3 describes a web-based implementation of the QualitySNP algorithm, called HaploSNPer, which is a tool for the detection of alleles and SNPs in user-specified input sequences from both diploid and polyploid species. HaploSNPer tries to find homologous sequences in user-specified sequence databases using a user-supplied seed sequence, or in a collection of input sequences. All alleles and associated SNPs are identified on clusters of these homologous sequences using QualitySNP. HaploSNPer provides a user-friendly interface for visualization of SNPs and alleles, which allows the selection of informative SNPs and allele specific makers. Currently HaploSNPer is available for nine animal and thirteen plant species, and is available from http://www.bioinformatics.nl/tools/haplosnper/ . Chapter 4 presents a new tool, called PolySSR, to identify polymorphic SSRs rather than just SSRs based on public EST sequence data derived from heterozygotes and/or different genotypes. Based on PolySSR a pipeline was developed to automatically develop primers for putatively polymorphic SSR markers, taking into account SNPs in the SSR flanking regions, thus improving the success rate of the potential markers. Furthermore, SSR positions in coding or UTR regions of genes are identified by the pipeline. The pipeline also includes a searchable database for these SSRs. The value of PolySSR was demonstrated by the fact that nearly all tested SSRs predicted to be polymorphic were indeed validated as polymorphic, and also most designed primers produced clear amplicons. Large numbers of polymorphic SSRs were identified from publicly available ESTs in potato, tomato, rice, Arabidopsis, brassica and chicken using the pipeline. They are stored into a database, which is available at http://www.bioinformatics.nl/tools/polyssr/ . PolySSR not only decreases the cost of designing and testing primers, it also brings a new approach to use the redundancy and heterozygosity of ESTs for developing SSRs that was ignored before. Analysis of the data obtained with polySSR showed that a larger percentage of short SSRs identified in the species used in the study were polymorphic than that of long SSRs. From this it is clear that in the past we have ‘forgotten’ a whole class of putatively informative markers. QualitySNP and PolySSR have been implemented into a pipeline called GeneTagger to find candidate genes underlying a QTL using the strategy of narrowing the QTL interval (Chapter 5). The pipeline first detects the syntenic region of a QTL interval in the species under study in a model species based on marker sequences linked to the QTL. Next, within the syntenic regions identified in the model species genes are identified that might have a function related to the QTL or genes within the regions that are not part of a large gene family. Based on their map position in the model species a number of genes are selected for marker development. To facilitate marker development, ESTs derived from the target species are analyzed using QualitySNP, PolySSR and other tools. Based on identified genetic variations in the selected genes, molecular markers can be developed for accurate fine-mapping of the QTL and ultimately identification of the gene underlying the QTL effect. The pipeline has been used to narrow the clubroot resistance QTL. The tool is available from the website http://www.bioinformatics.nl/tools/genetagger/ . Finally, merits and shortcomings of the tools that have been developed as well as related bioinformatics questions that arose during these studies are discussed in Chapter 6.

The thesis focuses on two bioinformatics research topics: the development of tools for an efficient and reliable identification of single nucleotides polymorphisms (SNPs) and polymorphic simple sequence repeats (SSRs) from expressed sequence tags (ESTs) (Chapter 2, 3 and 4), and the subsequent implementation of these tools in a pipeline for narrowing down QTL intervals to facilitate the identification of candidate genes for the QTL (Chapter 5). Chapter 1 provides an introduction to molecular markers, SNP and SSR markers, illustrated by a number of applications of molecular markers, as well as existing programs for their detection. After analysis of existing problems and programs for the detection SNPs, a new algorithm is described to reliably identify SNPs and indels in EST data from diploid and polyploid species (Chapter 2). The algorithm is implemented in a program called QualitySNP. This program uses three filters to identify reliable SNPs: filter one screens for potential SNPs by requiring at least two sequences per represented allele; filter two uses a haplotype-based strategy to filter out clusters with paralogs and false SNPs caused by sequence errors; finally, filter three calculates a confidence score for every putative SNP according to the number of occurrences of each allele in high and low quality regions. For the detection of non-synonymous SNPs (nsSNP), synonymous SNPs and SNPs in UTRs, a program was developed as well. Furthermore, these programs were implemented in a pipeline that includes the identification of SNPs and nsSNPs, as well as a storage and retrieval system. Using this pipeline, large numbers of SNPs could be identified in potato ESTs. QualitySNP is available for running on LINUX and UNIX systems; the program, user manual and examples are available at http://www.bioinformatics.nl/tools/snpweb/ . Chapter 3 describes a web-based implementation of the QualitySNP algorithm, called HaploSNPer, which is a tool for the detection of alleles and SNPs in user-specified input sequences from both diploid and polyploid species. HaploSNPer tries to find homologous sequences in user-specified sequence databases using a user-supplied seed sequence, or in a collection of input sequences. All alleles and associated SNPs are identified on clusters of these homologous sequences using QualitySNP. HaploSNPer provides a user-friendly interface for visualization of SNPs and alleles, which allows the selection of informative SNPs and allele specific makers. Currently HaploSNPer is available for nine animal and thirteen plant species, and is available from http://www.bioinformatics.nl/tools/haplosnper/ . Chapter 4 presents a new tool, called PolySSR, to identify polymorphic SSRs rather than just SSRs based on public EST sequence data derived from heterozygotes and/or different genotypes. Based on PolySSR a pipeline was developed to automatically develop primers for putatively polymorphic SSR markers, taking into account SNPs in the SSR flanking regions, thus improving the success rate of the potential markers. Furthermore, SSR positions in coding or UTR regions of genes are identified by the pipeline. The pipeline also includes a searchable database for these SSRs. The value of PolySSR was demonstrated by the fact that nearly all tested SSRs predicted to be polymorphic were indeed validated as polymorphic, and also most designed primers produced clear amplicons. Large numbers of polymorphic SSRs were identified from publicly available ESTs in potato, tomato, rice, Arabidopsis, brassica and chicken using the pipeline. They are stored into a database, which is available at http://www.bioinformatics.nl/tools/polyssr/ . PolySSR not only decreases the cost of designing and testing primers, it also brings a new approach to use the redundancy and heterozygosity of ESTs for developing SSRs that was ignored before. Analysis of the data obtained with polySSR showed that a larger percentage of short SSRs identified in the species used in the study were polymorphic than that of long SSRs. From this it is clear that in the past we have ‘forgotten’ a whole class of putatively informative markers. QualitySNP and PolySSR have been implemented into a pipeline called GeneTagger to find candidate genes underlying a QTL using the strategy of narrowing the QTL interval (Chapter 5). The pipeline first detects the syntenic region of a QTL interval in the species under study in a model species based on marker sequences linked to the QTL. Next, within the syntenic regions identified in the model species genes are identified that might have a function related to the QTL or genes within the regions that are not part of a large gene family. Based on their map position in the model species a number of genes are selected for marker development. To facilitate marker development, ESTs derived from the target species are analyzed using QualitySNP, PolySSR and other tools. Based on identified genetic variations in the selected genes, molecular markers can be developed for accurate fine-mapping of the QTL and ultimately identification of the gene underlying the QTL effect. The pipeline has been used to narrow the clubroot resistance QTL. The tool is available from the website http://www.bioinformatics.nl/tools/genetagger/ . Finally, merits and shortcomings of the tools that have been developed as well as related bioinformatics questions that arose during these studies are discussed in Chapter 6.

Diversity Arrays Technology (DArT) is a microarray-based DNA marker technique for genome-wide discovery and genotyping of genetic variation. DArT allows simultaneous scoring of hundreds- to thousands of restriction site based polymorphisms between genotypes and does not require DNA sequence information or site-specific oligonucleotides. In contrast to many other gel-based marker technologies DArT markers can be sequenced readily. Chapter 1 gives an overview of the most widely used genetic marker methods and their limitations. In addition some background information is provided on the wheat infecting fungal pathogen Mycosphaerella graminicola , in which we applied DArT.In Chapter 2 the DArT procedure is described in more detail. All steps in the DArT protocol are explained and discussed. In addition some adaptations to the DArT procedure with respect to the complexity reduction and use of adapter sequences are demonstrated on the basis of experiments performed in M. graminicola . Finally, the specifically developed software (DArTsoft) and the application of DArT markers in the mapping process are described.Chapter 3demonstrates the potential of DArT for genetic mapping by validating the quality and molecular basis of the markers, using the model plant Arabidopsis thaliana . Restriction fragments from a genomic representation of the ecotype Landsberg erecta (L er ) were amplified by PCR, individualized by cloning and spotted onto glass slides. The arrays were then hybridized with labeled genomic representations of the ecotypesColumbia(Col) and L er and of individuals from an F 2 population obtained from aCol× L er cross. The scoring of markers with specialized software was highly reproducible and 107 markers could unambiguously be ordered on a genetic linkage map. The marker order on the genetic linkage map coincided with the order on the DNA sequence map. Sequencing of the L er markers and alignment with the availableColgenome sequence confirmed that the polymorphism in DArT markers is largely a result of restriction site polymorphisms.In Chapter 4DArT was used to construct two high-density genetic linkage maps of the haploid wheat pathogen M. graminicola . The isolate IPO323 that originates from a bread wheat field was crossed in planta with either another bread wheat isolate IPO94269 or the durum wheat isolate IPO95052. Host species specificity of the plant pathogenic fungus M. graminicola towards hexaploid bread wheat and tetraploid durum wheat is well-documented. Although M. graminicola has an important sexual cycle and in some parts of the world durum wheat and bread wheat are grown side by side, in nature most isolates are pathogenic to either bread wheat or durum wheat. The generation of progeny from the IPO323 x IPO95052 in planta cross was therefore unexpected as these parental isolates are avirulent to durum wheat and bread wheat, respectively. The segregation of markers was followed in the progenies of both crosses. In total 2078 markers (of which 1793 were DArT markers) could be mapped. The maps of the individual crosses showed absolute co-linearity to a constructed bridge map, with the exception of eight markers that appeared to reveal translocations between IPO95052 and IPO94269. Graphical genotyping enabled the identification of dispensable chromosomes in 15-20% of the progenies, although these chromosomes were present in both parents. We demonstrate that this is due to aberrations (partly by non-disjunction) during meiosis. The loss of chromosomes did not seem to affect viability and pathogenicity. This is the first report on detailed meiotic events using high-density mapping in a haploid filamentous fungus. We show that DArT is a highly efficient and robust marker technology for genetic mapping. The DArT markers can be readily sequenced and are currently being deployed in the assembly of the 9X M. graminicola IPO323 genome sequence.In Chapter 5 the genetic linkage map of the IPO323 x IPO95052 cross was used to map several quantitative trait loci (QTL) for host species specificity in M. graminicola .Virulence of the progeny (n=163) was tested on four bread wheat cultivars and three durum wheat cultivars. All progeny isolates grew normally in vitro,but in contrast to crosses performed with two bread wheat isolates, nearly 50% were unable to cause disease on any of the generally susceptible bread wheat and durum wheat cultivars. Interestingly, other progeny caused disease on either bread wheat cultivars or durum wheat cultivars or both wheat species. This proves that the observed host species specificity was overcome in a single generation. In addition, some progeny were able to cause disease on the bread wheat cv. Shafir even though both parents are avirulent to this cultivar. Detailed analyses allowed the identification of QTLs that determine specificity using the high-density genetic linkage map containing 1144 molecular markers, described in the previous chapter. In total, we identified nine QTLs on seven linkage groups. Remarkably, no specific locus was identified that explains the host species specificity to either durum or bread wheat. One locus with a very high LOD value had a major effect on specificity towards all tested durum wheat cultivars, but was previously identified as a locus involved in cultivar specificity. To our knowledge, this is the first comprehensive QTL analysis on virulence performed on a plant pathogenic fungus. We conclude that specificity in M. graminicola is controlled by many genes and that the long-accepted notion of species specificity among isolates of M. graminicola may in fact be a mix of genetically inherited factors, whereby the distinction between host species specificity and cultivar specificity is not clear-cut.Chapter 6shows the alignment of the DArT markers to the genome of M. graminicola that recently was sequenced by the US Department of Energy-Joint Genome Institute (DOE-JGI). Assembly (version 1.0) of the 9X shotgun sequence resulted in the identification of 129 scaffolds ranging in size from 99 kb to 6.36 Mbp and covering 41.9 Mbp. In this assembly a total of 36 telomeres were identified. Sequencing of DArT markers used for the construction of two high-density genetic linkage maps in M. graminicola provided us the opportunity to align the genetic maps with the physical (sequence) map. By performing BLAST and BLAT analysis of the sequences obtained, 1069 IPO323, 724 IPO95052 and 24 SSR markers could be positioned on the 22 largest scaffolds in the assembly allowing the most comprehensive alignment among sequenced fungal genomes. Comparison of the positions of these markers on the sequence map with the order of the markers on the genetic linkage maps resulted in a nearly complete co-linearity of the order of markers, covering >99 % of the current assembly. When a 2 cM confidence interval was applied, 99% of the markers were co-linear with the genome sequence. The alignment provided evidence that scaffolds 7 and 17, 10 and 14 and also 12 and 22 belong to one chromosome and that scaffolds 4 and 9 were wrongly assembled in the first draft. Furthermore, we identified large differences in genetic versus physical distance in certain genomic regions demonstrating that recombination frequencies increase towards the telomeres. In addition, we identified significant differences in the recombination rates in the different crosses studied. The integration of the genetic maps with the physical map will: 1) assist in finishing the genome assembly; 2) facilitate map-based cloning strategies of genes involved in pathogenesis and 3) help to understand the processes that occur during meiosis and recombination.Finally, Chapter 7 gives a general discussion. The advantages and disadvantages of the DArT marker method as well as alternative array-based technologies are discussed. We conclude that DArT is a cost effective marker technology which generates the extremely high quality data needed for high-density mapping. In addition the ease in which sequences of the markers are obtained, allow large-scale alignments between genetic and physical maps. This makes DArT the method of choice especially for species for which large genetic variation exists, limited financial resources are available or for complex polyploid genomes that may not be amenable to the whole-genome sequence approach. In addition the results obtained using the high-density genetic linkage maps from M. graminicola and the biological implications for the fungal population in respect to adaptation to the host are discussed.

Diversity Arrays Technology (DArT) is a microarray-based DNA marker technique for genome-wide discovery and genotyping of genetic variation. DArT allows simultaneous scoring of hundreds- to thousands of restriction site based polymorphisms between genotypes and does not require DNA sequence information or site-specific oligonucleotides. In contrast to many other gel-based marker technologies DArT markers can be sequenced readily. Chapter 1 gives an overview of the most widely used genetic marker methods and their limitations. In addition some background information is provided on the wheat infecting fungal pathogen Mycosphaerella graminicola , in which we applied DArT.In Chapter 2 the DArT procedure is described in more detail. All steps in the DArT protocol are explained and discussed. In addition some adaptations to the DArT procedure with respect to the complexity reduction and use of adapter sequences are demonstrated on the basis of experiments performed in M. graminicola . Finally, the specifically developed software (DArTsoft) and the application of DArT markers in the mapping process are described.Chapter 3demonstrates the potential of DArT for genetic mapping by validating the quality and molecular basis of the markers, using the model plant Arabidopsis thaliana . Restriction fragments from a genomic representation of the ecotype Landsberg erecta (L er ) were amplified by PCR, individualized by cloning and spotted onto glass slides. The arrays were then hybridized with labeled genomic representations of the ecotypesColumbia(Col) and L er and of individuals from an F 2 population obtained from aCol× L er cross. The scoring of markers with specialized software was highly reproducible and 107 markers could unambiguously be ordered on a genetic linkage map. The marker order on the genetic linkage map coincided with the order on the DNA sequence map. Sequencing of the L er markers and alignment with the availableColgenome sequence confirmed that the polymorphism in DArT markers is largely a result of restriction site polymorphisms.In Chapter 4DArT was used to construct two high-density genetic linkage maps of the haploid wheat pathogen M. graminicola . The isolate IPO323 that originates from a bread wheat field was crossed in planta with either another bread wheat isolate IPO94269 or the durum wheat isolate IPO95052. Host species specificity of the plant pathogenic fungus M. graminicola towards hexaploid bread wheat and tetraploid durum wheat is well-documented. Although M. graminicola has an important sexual cycle and in some parts of the world durum wheat and bread wheat are grown side by side, in nature most isolates are pathogenic to either bread wheat or durum wheat. The generation of progeny from the IPO323 x IPO95052 in planta cross was therefore unexpected as these parental isolates are avirulent to durum wheat and bread wheat, respectively. The segregation of markers was followed in the progenies of both crosses. In total 2078 markers (of which 1793 were DArT markers) could be mapped. The maps of the individual crosses showed absolute co-linearity to a constructed bridge map, with the exception of eight markers that appeared to reveal translocations between IPO95052 and IPO94269. Graphical genotyping enabled the identification of dispensable chromosomes in 15-20% of the progenies, although these chromosomes were present in both parents. We demonstrate that this is due to aberrations (partly by non-disjunction) during meiosis. The loss of chromosomes did not seem to affect viability and pathogenicity. This is the first report on detailed meiotic events using high-density mapping in a haploid filamentous fungus. We show that DArT is a highly efficient and robust marker technology for genetic mapping. The DArT markers can be readily sequenced and are currently being deployed in the assembly of the 9X M. graminicola IPO323 genome sequence.In Chapter 5 the genetic linkage map of the IPO323 x IPO95052 cross was used to map several quantitative trait loci (QTL) for host species specificity in M. graminicola .Virulence of the progeny (n=163) was tested on four bread wheat cultivars and three durum wheat cultivars. All progeny isolates grew normally in vitro,but in contrast to crosses performed with two bread wheat isolates, nearly 50% were unable to cause disease on any of the generally susceptible bread wheat and durum wheat cultivars. Interestingly, other progeny caused disease on either bread wheat cultivars or durum wheat cultivars or both wheat species. This proves that the observed host species specificity was overcome in a single generation. In addition, some progeny were able to cause disease on the bread wheat cv. Shafir even though both parents are avirulent to this cultivar. Detailed analyses allowed the identification of QTLs that determine specificity using the high-density genetic linkage map containing 1144 molecular markers, described in the previous chapter. In total, we identified nine QTLs on seven linkage groups. Remarkably, no specific locus was identified that explains the host species specificity to either durum or bread wheat. One locus with a very high LOD value had a major effect on specificity towards all tested durum wheat cultivars, but was previously identified as a locus involved in cultivar specificity. To our knowledge, this is the first comprehensive QTL analysis on virulence performed on a plant pathogenic fungus. We conclude that specificity in M. graminicola is controlled by many genes and that the long-accepted notion of species specificity among isolates of M. graminicola may in fact be a mix of genetically inherited factors, whereby the distinction between host species specificity and cultivar specificity is not clear-cut.Chapter 6shows the alignment of the DArT markers to the genome of M. graminicola that recently was sequenced by the US Department of Energy-Joint Genome Institute (DOE-JGI). Assembly (version 1.0) of the 9X shotgun sequence resulted in the identification of 129 scaffolds ranging in size from 99 kb to 6.36 Mbp and covering 41.9 Mbp. In this assembly a total of 36 telomeres were identified. Sequencing of DArT markers used for the construction of two high-density genetic linkage maps in M. graminicola provided us the opportunity to align the genetic maps with the physical (sequence) map. By performing BLAST and BLAT analysis of the sequences obtained, 1069 IPO323, 724 IPO95052 and 24 SSR markers could be positioned on the 22 largest scaffolds in the assembly allowing the most comprehensive alignment among sequenced fungal genomes. Comparison of the positions of these markers on the sequence map with the order of the markers on the genetic linkage maps resulted in a nearly complete co-linearity of the order of markers, covering >99 % of the current assembly. When a 2 cM confidence interval was applied, 99% of the markers were co-linear with the genome sequence. The alignment provided evidence that scaffolds 7 and 17, 10 and 14 and also 12 and 22 belong to one chromosome and that scaffolds 4 and 9 were wrongly assembled in the first draft. Furthermore, we identified large differences in genetic versus physical distance in certain genomic regions demonstrating that recombination frequencies increase towards the telomeres. In addition, we identified significant differences in the recombination rates in the different crosses studied. The integration of the genetic maps with the physical map will: 1) assist in finishing the genome assembly; 2) facilitate map-based cloning strategies of genes involved in pathogenesis and 3) help to understand the processes that occur during meiosis and recombination.Finally, Chapter 7 gives a general discussion. The advantages and disadvantages of the DArT marker method as well as alternative array-based technologies are discussed. We conclude that DArT is a cost effective marker technology which generates the extremely high quality data needed for high-density mapping. In addition the ease in which sequences of the markers are obtained, allow large-scale alignments between genetic and physical maps. This makes DArT the method of choice especially for species for which large genetic variation exists, limited financial resources are available or for complex polyploid genomes that may not be amenable to the whole-genome sequence approach. In addition the results obtained using the high-density genetic linkage maps from M. graminicola and the biological implications for the fungal population in respect to adaptation to the host are discussed.