Dissertations / Theses on the topic 'Phenotype'

Naturally evolving proteins must navigate a vast set of possible sequences to evolve new functions. This process depends on the genotype-phenotype map. Much effort has been directed at measuring protein genotype phenotype maps to uncover evolutionary trajectories that lead to new functions. Often, these maps are too large to comprehensively measure. Sparsely measured maps, however, are prone to missing key evolutionary trajectories. Many groups turn to computational models to infer missing phenotypes. These models treat mutations as independent perturbations to the genotype-phenotype map. A key question is how to handle non-independent effects known as epistasis. In this dissertation, we address two sources of epistasis: 1) global and 2) local epistasis. We find that incorporating global epistasis improves our predictive power, while local epistasis does not. We use our model to infer unknown phenotypes in the Plasmodium falciparum chloroquine transporter (PfCRT) genotype-phenotype map, a protein responsible for conferring drug resistance in malaria. From these predictions, we uncover key evolutionary trajectories that led high resistance. This dissertation includes previously published and unpublished co-authored material.
2020-01-11

Introduction: The purpose of this study was to examine phenotypic expression of craniofacial form, shape, and size as it relates to the genotype of an individual. Shape analyses were completed on 3-D images of each subject's craniofacial structure by landmarking 45 points of interest on the cranial base, facial bones, and upper and lower jaws. A candidate gene analysis was undertaken focusing on specific genes in the Hippo Signaling Pathway to examine genotype-phenotype correlations that play a role in craniofacial development. This study is a continuation of a larger project aimed at the identification of candidate genes associated with human dento-skeletal bite problems led by Dr. Lina Moreno-Uribe. Methods: The sample size for our study included 166 individuals who had never been treated orthodontically at the time of records. Each individual was genotyped and a CBCT of the craniofacial complex was captured. Each CBCT image was landmarked by a single observer using 45 points to mark points on the cranial base and facial bones including the maxilla and mandible. General Procrustes superimposition was used to find correlations with phenotype and genotype. Size analysis was completed with average Euclidean Distances and ANOVA analysis. Results: 2 SNP's from the FOX03 gene had significant associations with size. The AA genotyped individuals appeared larger in overall size than AB genotyped individuals. 3 SNP's had statistically significant associations with facial form. The FOX06 SNPs had significant associations with increased anterior-posterior growth of the maxilla. The AJUBA SNP had significant associations with increased overall craniofacial breadth. Conclusion: Genes in the Hippo signaling pathway have specific roles in the development of facial form and size.

Thioredoxin is a small ubiquitous oxido-reductase found in all species. The highly conserved active site, which facilitates thioredoxins redox activity, contains two redox active cysteine residues. Thioredoxin has numerous protein substrates to which it donates H+ ions and it can also function as a free radical scavenger. Through these activities thioredoxin is able to influence the redox state of not only its protein targets, but also the entire cellular environment. Thioredoxin has been implicated in many biological functions, and one mechanism by which it influences these functions is through interactions with a number of transcription factors including NF-kappa-B and p53. Thioredoxin also has numerous extracellular biological roles. It has been shown that thioredoxin is actively secreted from a number of normal and transformed cell lines including fibroblasts and activated B and T cells. This study investigates the role of thioredoxin in embryonic implantation and cancer cell metastasis, two physiological functions which rely on the same basic processes. Thioredoxin expression has previously been shown to be increased in many cancers. However it has not yet been established whether this increase is a causative or a side effect of the cancerous phenotype. Similarly thioredoxin expression has previously been shown to be increased during different phases of the oestrus cycle and pregnancy. This thesis describes the role of thioredoxin in embryonic implantation using a marmoset model. A thioredoxin cDNA was isolated and subsequently sequenced. Preliminary antibody experiments indicated that the anti human thioredoxin monoclonal antibodies available in our laboratory would recognise marmoset thioredoxin. Subsequently immunocytochemistry using anti human thioredoxin antibodies was carried out on sectioned marmoset uterus and embryonic tissue. The results indicated that thioredoxin is expressed by cells at the embryonic-maternal interface of early implantation sites. Further studies demonstrated that thioredoxin is also expressed and secreted by cultured blastocysts in vitro. This thesis also describes the role of thioredoxin in cancer cell metastasis. Results of this study indicate that thioredoxin is actively involved in facilitating the invasive phenotype of breast cancer cells. The two cell lines utilised were MCF-7, a well differentiated, relatively non-invasive breast cancer cell line; and MDA-MB-231, a poorly differentiated, highly invasive breast cancer cell line. The cell lines were transfected with thioredoxin sense, antisense and 1SS (encodes thioredoxin with both active cysteine residues mutated to serine residues and is thus redox inactive) constructs. The results demonstrate that when endogenous thioredoxin levels are increased, i.e. transfected with a sense thioredoxin construct, the invasive breast cancer cell line MDA-MB-231 becomes more invasive, conversely when endogenous levels are decreased, i.e. transfected with antisense or 1SS constructs, the invasive capacity of these cells decreases. However, when the endogenous level of thioredoxin was manipulated in the relatively non-invasive cell line MCF-7 very little effect was observed. Results also indicate that thioredoxin has the ability to act as a chemoattractant for actively invading breast cancer cells. Both of these functions appear to be dependent on thioredoxin's redox activity. Additional studies described in this thesis have shown that thioredoxin is involved in the regulation of Sp1 in vitro. Sp1 is a transcription factor known to regulate the transcription of a number of genes whose products are intimately involved in the invasive phenotype. The results in this study suggest that Sp1 DNA binding is regulated by thioredoxin such that when reduced by the enzyme its binding to DNA is facilitated. Results also indicate that Sp1 may regulate the transcription of thioredoxin by binding to Sp1 sites within the thioredoxin promoter.

Thioredoxin is a small ubiquitous oxido-reductase found in all species. The highly conserved active site, which facilitates thioredoxins redox activity, contains two redox active cysteine residues. Thioredoxin has numerous protein substrates to which it donates H+ ions and it can also function as a free radical scavenger. Through these activities thioredoxin is able to influence the redox state of not only its protein targets, but also the entire cellular environment. Thioredoxin has been implicated in many biological functions, and one mechanism by which it influences these functions is through interactions with a number of transcription factors including NF-_B and p53. Thioredoxin also has numerous extracellular biological roles. It has been shown that thioredoxin is actively secreted from a number of normal and transformed cell lines including fibroblasts and activated B and T cells. This study investigates the role of thioredoxin in embryonic implantation and cancer cell metastasis, two physiological functions which rely on the same basic processes. Thioredoxin expression has previously been shown to be increased in many cancers. However it has not yet been established whether this increase is a causative or a side effect of the cancerous phenotype. Similarly thioredoxin expression has previously been shown to be increased during different phases of the oestrus cycle and pregnancy. This thesis describes the role of thioredoxin in embryonic implantation using a marmoset model. A thioredoxin cDNA was isolated and subsequently sequenced. Preliminary antibody experiments indicated that the anti human thioredoxin monoclonal antibodies available in our laboratory would recognise marmoset thioredoxin. Subsequently immunocytochemistry using anti human thioredoxin antibodies was carried out on sectioned marmoset uterus and embryonic tissue. The results indicated that thioredoxin is expressed by cells at the embryonic-maternal interface of early implantation sites. Further studies demonstrated that thioredoxin is also expressed and secreted by cultured blastocysts in vitro. This thesis also describes the role of thioredoxin in cancer cell metastasis. Results of this study indicate that thioredoxin is actively involved in facilitating the invasive phenotype of breast cancer cells. The two cell lines utilised were MCF-7, a well differentiated, relatively non-invasive breast cancer cell line; and MDA-MB-231, a poorly differentiated, highly invasive breast cancer cell line. The cell lines were transfected with thioredoxin sense, antisense and 1SS (encodes thioredoxin with both active cysteine residues mutated to serine residues and is thus redox inactive) constructs. The results demonstrate that when endogenous thioredoxin levels are increased, i.e. transfected with a sense thioredoxin construct, the invasive breast cancer cell line MDA-MB-231 becomes more invasive, conversely when endogenous levels are decreased, i.e. transfected with antisense or 1SS constructs, the invasive capacity of these cells decreases. However, when the endogenous level of thioredoxin was manipulated in the relatively non-invasive cell line MCF-7 very little effect was observed. Results also indicate that thioredoxin has the ability to act as a chemoattractant for actively invading breast cancer cells. Both of these functions appear to be dependent on thioredoxin's redox activity. Additional studies described in this thesis have shown that thioredoxin is involved in the regulation of Sp1 in vitro. Sp1 is a transcription factor known to regulate the transcription of a number of genes whose products are intimately involved in the invasive phenotype. The results in this study suggest that Sp1 DNA binding is regulated by thioredoxin such that when reduced by the enzyme its binding to DNA is facilitated. Results also indicate that Sp1 may regulate the transcription of thioredoxin by binding to Sp1 sites within the thioredoxin promoter.
Thesis (PhD Doctorate)
Doctor of Philosophy (PhD)
School of Biomolecular and Biomedical Sciences
Full Text

Le malattie mitocondriali (MD) sono un gruppo clinicamente eterogeneo di patologie dovute al difetto di funzione dei mitocondri, in particolare della catena respiratoria mitocondriale (RC) e della fosforilazione ossidativa (OXPHOS). Le funzioni mitocondriali sono sotto il controllo di due diversi genomi: DNA mitocondriale (mtDNA) e genoma nucleare (nDNA). Le forme infantili sono più spesso associate a mutazioni nel nDNA; negli ultimi anni tecnologie di sequenziamento di nuova generazione (Next Generation Sequencing-NGS) hanno permesso di identificare nuovi geni malattia; in collaborazione con altri centri abbiamo contribuito alla definizione del fenotipo associato ai nuovi geni malattia identificati. L’applicazione di tale tecnica è stata estesa anche allo studio del mtDNA: anche se sono state descritte più di 100 mutazioni e delezioni nel mtDNA in associazione a uno spettro estremamente eterogeneo di presentazioni cliniche, solo alcune di esse sono associate a sindromi cliniche ben definite nell'infanzia. Abbiamo effettuato una valutazione sistematica dei dati clinici, strumentali, metabolici e biochimici di una vasta coorte di pazienti affetti dal MD più comune nell'infanzia: la sindrome di Leigh. Abbiamo analizzato in questa popolazione la correlazione genotipo-fenotipo nei casi associati al gene nDNA e mtDNA al fine di identificare indizi diagnostici per la sindrome di Leigh correlata al mtDNA. In casi geneticamente irrisolti e vari fenotipi (sindrome di Leigh, leucodistrofia ...), abbiamo eseguito lo screening del mtDNA utilizzando tecnologie di sequenziamento di nuova generazione (NGS) al fine di valutare, con elevata precisione, mutazioni puntiformi e delezioni singole o multiple di grandi dimensioni, entrambe in omoplasma o stato eteroplasmico. Abbiamo identificato mutazioni sia nuove che note associate a fenotipo inatteso (es. MTCO3). I nostri dati definiscono e ampliano meglio lo spettro fenotipico del mtDNA-MD nell'infanzia.
Mitochondrial diseases (MD) are a clinically heterogeneous group of disorders that arise as a result of dysfunction of the mitochondrial respiratory chain (RC) and oxidative phosphorylation (OXPHOS). Mitochondrial functions are under the control of two different genomes: mitochondrial DNA (mtDNA) and nuclear genome (nDNA). Childhood phenotypes are often associated with nDNA mutations; in recent years, new-generation sequencing technologies (Next Generation Sequencing-NGS) have identified novel causative genes; in collaboration with other centers we contributed to the definition of phenotype associated with the new identified disease genes. The application of this technique has also been extended to the study of mtDNA: even if more than 100 mutations and deletions in mtDNA have been described in association with an extremely heterogeneous spectrum of clinical presentations, only a few of them are associated with well-defined clinical syndromes in childhood. We performed a systematic evaluation of clinical, instrumental, metabolic and biochemical data of a large cohort of patients affected by the most common MD in childhood: Leigh syndrome. We analyzed in this population, genotype-phenotype correlation in nDNA and mtDNA gene associated cases in order to identify diagnostic clues for mtDNA related Leigh syndrome. In genetically unresolved cases and various phenotypes (Leigh syndrome, leukodystropy..), we performed mtDNA screening using next-generation sequencing (NGS) technologies in order to assess, with high accuracy, point mutations and single or multiple large deletions, both in homoplasmic or heteroplasmic state. We identified both novel and known mutations associated to unexpected phenotype (i.e. CO3 gene). Our data better define and expand the phenotypic spectrum of mtDNA-MD in childhood.

Breastfeeding in infancy, when compared with formula feeding, is associated with a reduced incidence of components of the metabolic syndrome later in life. One potential mechanism is via an effect on lipid metabolism and storage, manifesting as altered adiposity and ectopic lipid deposition. I have examined the null hypothesis: no association is detectable between infant feeding and adiposity or ectopic lipid in infancy, through a meta-analysis of published studies and a prospective cohort study of healthy infants employing gold standard direct measurement techniques (magnetic resonance imaging and spectroscopy). Eleven studies were identified for meta-analysis: in formula-fed compared to breastfed infants, fat mass was lower at 3-4 months [mean difference (95% confidence interval)]: [-0.09 kg (-0.18, -0.01 kg)] and 6 months [-0.18 kg (- 0.34, -0.01 kg)]. Conversely, at 12 months, fat mass was higher in formula-fed infants [0.29 kg (-0.03, 0.61 kg)] than in breastfed infants. Eighty-seven infants were included in a prospective cohort, of which 73 were investigated at two time points. In healthy, term, breastfed infants adipose tissue accretion between birth and 2-3 months ages was predominantly within subcutaneous rather than internal adipose tissue compartments, and a significant increase in intrahepatocellular lipid was detected: median [interquartile range] 0.653 [0.367-1.900] after birth and 1.837 [1.408-2.429] at 2-3 months. Comparing breastfed with formula fed infants within this cohort no significant differences were detected in total adipose tissue, adipose tissue distribution or intrahepatocellular lipid between birth and 2-3 months. Significant associations were detected between maternal BMI, rate of weight gain in early infancy and gender, and adipose tissue partitioning at 2-3 months. While method of feeding is associated with altered infant fat mass up to 6 months, no association is detectable with adipose tissue partitioning or ectopic hepatic lipid at 2-3 months.

Consistency of aggressiveness within and between situations was examined using a repeated social challenge test (Resident Intruder Test) at various key stages during the lifetime of male and female growing (meat) pigs and a subset of female breeding pigs. Behaviour during mixing at weaning and as gilts was examined and information was collected about the way in which these animals behaved in a social context, including measures related to, but not directly correlated with aggression, such as social status. The effects of age/maturity and experience were also considered. To understand how aggression might integrate within personality the various social measures were compared with cortisol (as a physiological indicator of stress) an a challenging situation unrelated to social confrontation; maternal behaviour was chosen as it is commercially relevant and has important implications for the welfare of gilts and piglets. Pigs were consistent in their responses to the RIT, but there were differences between sexes. Aggressiveness was consistent over a long period of time in female pigs, even with a gap of 90 days between tests and the onset of puberty. Male pigs showed an unexpectedly high level of mounting behaviour from a young age, which increased with maturity. Experience of the RIT improved consistency of responses and age at first testing affected both the speed of attacking and occurrence of attacks; those pigs experiencing the test when younger were more likely to and quicker to attach. RIT aggressiveness was however, not predictive of subsequent aggressiveness at mixing. As with the RIT, there were clear sex-differences observed during mixing at weaning with males being more aggressive, more successful in fights, more likely to mount and less likely to play than females. Pigs employed different strategies during mixing, the extremes of which were categorised by high-play-low-aggressiveness and vice versa. As expected, aggressive individuals were involved in more fights and won more fights, but suffered more skin lesions than non-aggressive individuals. Pigs that engaged in high-playing were generally the least successful in fights, but suffered fewer lesions and had equal ultimate dominance rank to aggressive pigs. The structure of the mixes changes between weaning and puberty. Fighting ceased sooner during the gilt mix, but aggression was more frequent and more severe. Gilts that reacted aggressively to their piglets were more aggressive and successful in the mix and more ‘reactive’ during farrowing. There were other links between farrowing and mixing behaviour, such as more frequent posture changes but less frequent nesting with greater mix-aggressiveness; indicating that aggressiveness and maternal behaviour are linked through personality.

Thesis (Master of Architecture)--University of Cincinnati, 2007.
Title from electronic theses title page (viewed July 16, 2007). Includes abstract. Keywords: architecture; typology; therapeutic; riding stable Includes bibliographic references.

Transvection, a chromosome pairing-dependent form of trans-based gene regulation, is widespread in the Drosophila melanogaster genome. Recent studies demonstrate that transvection is sensitive to cell environment and type in D. melanogaster, implicating transvection as a complex trait. To test this possibility, we first established that trans-interactions previously documented at the Malic enzyme (Men) locus are transvection (i.e., pairing-dependent). We then characterized the sensitivity of transvection at the Men locus to changes in the environment (temperature) and genetic background (third chromosome). Transvection varied significantly across genetic backgrounds and was significantly reduced by changes in temperature, and the two factors interacted to further modify transvection, while cis-based gene regulation remained unchanged by temperature. To determine if differences in transvection observed across genetic background and temperature are related to their effects on transcription factor expression, and possibly the presence or absence of binding sites for these transcription factors within the Men locus, we tested the relationship between Men expression and five transcription factors with binding sites near the Men transcription start sit (TSS). We found correlations between the expression of at least one transcription factor, Abd-B, and the presence of binding sites for that factor, and Men expression across changes in the environment. We also determined that changes in Abd-B expression can directly affect Men expression in cis, suggesting that cis and trans-regulation can share regulatory components in at least some cases. Together, our findings stress the importance of studying genetic interactions from a dynamic perspective by incorporating both genetic and environmental variation.

Darier's Disease (DD) is a rare autosomal dominantly inherited skin disorder in which co-occurrence of neuropsychiatric abnormalities has been frequently reported by dermatologists. The disease is caused by mutations in a single gene, ATP2A2, which maps to 12q23-q24.1 and is expressed in the skin and brain. This gene encodes SERCA2 (sarco/endoplasmic reticulum Ca 2+ ATPase isoform 2), a calcium pump involved in intracellular calcium transport. This study aimed to conduct the first systematic investigation of the neuropsychiatric phenotype in DD, investigate possible genotype-phenotype correlations, and compare the neuropsychiatric features in DD individuals and their first-degree unaffected relatives. One hundred unrelated individuals with DD and 24 of their unaffected relatives were assessed using a battery of standardised neuropsychiatric measures. The relationship between the mutations detected in iheATP2A2 gene and the presence/severity of neuropsychiatric phenotypes was examined. DD individuals reported high rates of mood disorders, specifically major depression (30%), suicide attempts (13%) and suicidal thoughts (31%), and these were significantly more common in DD when compared with normative population data. Further, individuals with DD reported higher scores on measures of neuropsychiatric dysfunction than their unaffected relatives. These associations cannot be explained by psychosocial factors. Mutations found among individuals with similar neuropsychiatric phenotypes clustered in certain locations within the SERCA2b protein. Together, these findings support the hypothesis that mutations mATP2A2 confer susceptibility to neuropsychiatric dysfunction, in particular mood disorders, in individuals with DD. These findings highlight the need for assessment and recognition of psychiatric symptoms in DD. The findings may also have implications for identification of other genetic factors involved in conferring susceptibility to neuropsychiatric features in individuals without DD. Further research is needed into other neuropsychiatric phenotypes in DD and into the specific functional effects of mutations in ATP2A2 and the relationship of these to the presence of certain neuropsychiatric phenotypes.

The porphyrias are a group of disorder resulting from defects in the haem biosynthesis pathway. The majority of defects are genetic in origin. The clinical penetrance of porphyria is variable and cannot be explained by environmental factors alone. It is therefore likely that genetic modifiers are present to influence the phenotypic outcome of the disease. In the first part of this study, variants in ABCB6, a member of the ATP-Binding Cassette transporter, were detected in a high proportion (62.5%) of severely affected porphyria patients. In vitro functional studies revealed that these ABCB6 variants either affect dimer formation (p.Arg276Trp, p.Thr521Ser, p.Gly588Ser and p.Ala681Thr), or has decreased ATPase activity (p.Ala492Thr). In a double knock out mouse model deficiency in Fech (Fechm1Pas) and Abcb6 showed a more severe phenotype compared to Fechm1Pas-only mice. The results support that ABCB6 acts as a genetic modifier for porphyria patients. In the second part of this study, detailed study on a patient with fatal liver failure secondary to erythropoietic protoporphyria (EPP) revealed an atypical low expression FECH allele and a maternally inherited FECH p.Phe260Leu mutation giving rise to the EPP. Exome sequencing of this EPP patient and her parents revealed multiple functional variants and a novel 0.79Mb duplication, all involving genes in the bile secretion pathway. These variants affect both the basolateral re-uptake and canalicular secretion of bile acids as well as their constitution. In EPP, the excessive haem intermediate protoporphyrin IX is excreted in bile. The results here suggest that genetic variants in the bile secretion pathway may contribute to the risk of liver failure in EPP patients.

To unveil the biological complexity at the basis of the genotype-phenotype relation it is fundamental to integrate knowledge that is to integrate the different omics describing the levels of biological complexity: genomics, proteomics, transcriptomics, metabolomics and interactomics. The situation gets more complicated when we move the focus to diseases and phenotypes. The identification of molecular mechanisms behind different phenotypes offers a way to understand the processes that lead to disease insurgence and progression. Another issue in computational biology is the prediction of specific phenotypic effect of gene and protein variants, to test the performance of computational methods towards experiment in vivo and in vitro. The main aim of this thesis is to study the relations among genes, variations, diseases and phenotypes with the approaches of computational biology, integrating information from different resources to make a step forward in the direction of unveiling the biological complexity. After a general introduction, we present the webservers eDGAR and PhenPath, collecting and analysing the gene-disease associations and the phenotypes-biological processes associations, respectively. We then assessed whether disease-related variations induce perturbations of the protein stability. To this aim, we developed a new predictor called INPS-3D. We test our predictors participating in international experiments on specific study cases. Thanks to the expertise acquired in the field, we also collaborate with the Sant’Orsola Genetic Medical Unit of the Department of Medicine and Surgery of the University of Bologna, building a series of models of protein structure of myosin 1F and its variants related to the thyroid cancer. Concluding, we tried to depict the biological complexity merging a large-scale approach with the analysis of specific study cases, providing webservers, tools and computation methods to help researchers in directing further experiments.

La relació entre fenotip i informació genotípica és considerablement intricada i complexa. Els mètodes d’aprenentatge automàtic s’han utilitzat amb èxit per a la predicció de fenotips en un gran ventall de problemes dins de la genètica i la genòmica. Tanmateix, les dades biològiques sovent estan estructurades i són de tipus “no estàndard”, el que pot suposar un repte per a la majoria de mètodes d’aprenentatge automàtic. Entre aquestos, els mètodes kernel proporcionen un enfocament molt versàtil per manejar diferents tipus de dades i problemes mitjançant la utilització d’una família de funcions anomenades de kernel. L’objectiu principal d’aquesta tesi doctoral és el desenvolupament i l’avaluació d’estratègies de kernel específiques per a la predicció fenotípica, especialment en problemes biològics amb dades o dissenys experimentals de tipus estructurat. A la primera part, utilitzam seqüències de proteasa, transcriptasa inversa i integrasa per predir la resistència del VIH a fàrmacs antiretrovirals. Proposam dos kernels categòrics (Overlap i Jaccard) que tenen en compte les particularitats de les dades de VIH, com per exemple les barreges d’al·lels. Els kernels proposats es combinen amb Support Vector Machines (SVM) i es comparen amb dos kernels estàndard (Linear i RBF) i dos mètodes que no són de kernel: els boscos aleatoris (RF) i un tipus de xarxa neuronal (el perceptró multicapa). També incloem en els kernels la importància relativa de cada posició de la proteïna pel que fa a la resistència. Els resultats mostren que tenir en compte la naturalesa categòrica de les dades i la presència de barreges millora sistemàticament la predicció. L’efecte de ponderar les posicions per la seua importància és més gran en la transcriptasa inversa i en la integrasa, el que podria estar relacionat amb les diferències que hi ha entre els tres enzims pel que fa als patrons de mutació per adquirir resistència a fàrmacs antiretrovirals. A la segona part, ampliam l’estudi anterior per considerar no-independència entre les posicions de la proteïna. Representam les proteïnes com a grafs i ponderam cada aresta entre dos residus per la seua distància euclidiana, calculada a partir de dades de cristal·lografia de rajos X. A continuació, els aplicam un kernel per a grafs (el random walk exponential kernel) que integra els kernels Overlap i Jaccard. A pesar dels avantatges potencials d’aquest kernel, no aconseguim millorar els resultats obtinguts en la primera part. A la tercera part, proposam un kernel framework per unificar les anàlisis supervisades i no supervisades en el camp del microbioma. Aprofitam la mateixa matriu de kernel per predicció mitjançant SVM i visualització mitjançant anàlisi de components principals amb kernels (kPCA). Discutim com transformar mesures de beta-diversitat en kernels, i definim dos kernels per a dades composicionals (Aitchison-RBF i compositional linear). Aquest darrer kernel també permet obtenir les importàncies dels tàxons respecte del fenotip predit (signatures microbianes). Per a les dades amb estructuració espacial i temporal utilitzam Multiple Kernel Learning i kernels per a sèries temporals, respectivament. El framework s’il·lustra amb tres bases de dades: la primera conté mostres de sòl, la segona mostres humanes amb una component espacial i la tercera, no publicada fins ara, dades longitudinals de porcs. Totes les anàlisis es contrasten amb els estudis originals (en els dos primers casos) i també amb els resultats dels RF. El nostre kernel framework no només permet una visió global de les dades, sinó que també dóna bons resultats a cada àrea d’aprenentatge. En les anàlisis no supervisades, els patrons detectats en estudis previs es conserven a la kPCA. En anàlisis supervisades, el SVM té un rendiment superior (o equivalent) al dels RF, mentre que les signatures microbianes són coherents amb els estudis originals i la literatura prèvia.
La relación entre fenotipo e información genotípica es considerablemente intrincada y compleja. Los métodos de aprendizaje automático (ML) se han utilizado con éxito para la predicción de fenotipos en una gran variedad de problemas dentro de la genética y la genómica. Sin embargo, los datos biológicos suelen estar estructurados y pertenecer a tipos de datos "no estándar", lo que puede representar un desafío para la mayoría de los métodos de ML. Entre ellos, los métodos de kernel permiten un enfoque muy versátil para manejar diferentes tipos de datos y problemas mediante una familia de funciones llamadas de kernel. El objetivo principal de esta tesis doctoral es el desarrollo y evaluación de enfoques de kernel específicos para la predicción fenotípica, centrándose en problemas biológicos con tipos de datos o diseños experimentales estructurados. En la primera parte, usamos secuencias de proteínas mutadas del VIH (proteasa, transcriptasa inversa e integrasa) para predecir la resistencia a antiretrovirales. Proponemos dos funciones de kernel categóricas (Overlap y Jaccard) que tienen en cuenta las particularidades de los datos de VIH, como las mezclas de alelos. Los kernels propuestos se combinan con máquinas de vector soporte (SVM) y se comparan con dos funciones de kernel estándar (Linear y RBF) y dos métodos que no son de kernel: bosques aleatorios (RF) y un tipo de red neuronal, el perceptrón multicapa. También incluimos en los kernels la importancia relativa de cada posición de la proteína con respecto a la resistencia. Tener en cuenta tanto la naturaleza categórica de los datos como la presencia de mezclas obtenemos sistemáticamente mejores predicciones. El efecto de la ponderación es mayor en los inhibidores de la integrasa y la transcriptasa inversa, lo que puede estar relacionado con diferencias en los patrones mutacionales de las tres enzimas virales. En la segunda parte, ampliamos el estudio anterior para considerar que las posiciones de las proteínas pueden no ser independientes. Las secuencias mutadas se representan como grafos, ponderándose las aristas por la distancia euclidiana entre residuos obtenida por cristalografía de rayos X. A continuación, se calcula un kernel para grafos (el exponential random walk kernel) que integra los kernels Overlap y Jaccard. A pesar de las ventajas potenciales de este enfoque, no observamos una mejora en la capacidad predictiva. En la tercera parte, proponemos un kernel framework para unificar los análisis supervisados ​​y no supervisados del microbioma. Para ello, usamos una misma matriz de kernel para predecir fenotipos usando SVM y visualización a través de análisis de componentes principales con kernels (kPCA). Definimos dos kernels para datos composicionales (Aitchison-RBF y compositional linear) y discutimos la transformación de medidas de beta-diversidad en kernels. El kernel lineal composicional también permite la recuperación de importancias de taxones (firmas microbianas) del modelo SVM. Para datos con estructura espacial y temporal usamos Multiple Kernel Learning y kernels para series temporales, respectivamente. Ilustramos el kernel framework con tres conjuntos de datos: datos de suelos, datos humanos con un componente espacial y, un conjunto de datos longitudinales inéditos sobre producción porcina. Todos los análisis incluyen una comparación con los informes originales (en los dos primeros casos), así como un contraste con los resultados de RF. El kernel framework no solo permite una visión holística de los datos, sino que también da buenos resultados en cada área de aprendizaje. En análisis no supervisados, los principales patrones detectados en los estudios originales se conservan en kPCA. En análisis supervisados, la SVM tiene un rendimiento mayor (o equivalente) a los RF, mientras que las firmas microbianas son coherentes con los estudios originales y la literatura previa.
The relationship between phenotype and genotypic information is considerably intricate and complex. Machine Learning (ML) methods have been successfully used for phenotype prediction in a great range of problems within genetics and genomics. However, biological data is usually structured and belongs to & 'nonstandard' data types, which can pose a challenge to most ML methods. Among them, kernel methods bring along a very versatile approach to handle different types of data and problems through a family of functions called kernels. The main goal of this PhD thesis is the development and evaluation of specific kernel approaches for phenotypic prediction, focusing on biological problems with structured data types or study designs. In the first part, we predict drug resistance from HIV-mutated protein sequences (protease, reverse transcriptase and integrase). We propose two categorical kernel functions (Overlap and Jaccard) that take into account HIV data particularities, such as allele mixtures. The proposed kernels are coupled with Support Vector Machines (SVM) and compared against two well-known standard kernel functions (Linear and RBF) and two nonkernel methods: Random Forests (RF) and the Multilayer Perceptron neural network. We also include a relative weight into the aforementioned kernels, representing the importance of each protein residue regarding drug resistance. Taking into account both the categorical nature of data and the presence of mixtures consistently delivers better predictions. The weighting effect is higher in reverse transcriptase and integrase inhibitors, which may be related to the different mutational patterns in the viral enzymes regarding drug resistance. In the second part, we extend the previous study to consider the fact that protein positions are not independent. Mutated sequences are modeled as graphs, with edges weighted by the Euclidean distance between residues, obtained from crystal three-dimensional structures. A kernel for graphs (the exponential random walk kernel) that integrates the previous Overlap and Jaccard kernels is then computed. Despite the potential advantages of this kernel for graphs, an improvement on predictive ability as compared to the kernels of the first study is not observed. In the third part, we propose a kernel framework to unify unsupervised and supervised microbiome analyses. To do so, we use the same kernel matrix to perform phenotype prediction via SVMs and visualization via kernel Principal Components Analysis (kPCA). We define two kernels for compositional data (Aitchison-RBF and compositional linear) and discuss the transformation of beta-diversity measures into kernels. The compositional linear kernel also allows the retrieval of taxa importances (microbial signatures) from the SVM model. Spatial and time-structured datasets are handled with Multiple Kernel Learning and kernels for time series, respectively. We illustrate the kernel framework with three datasets: a single point soil dataset, a human dataset with a spatial component, and a previously unpublished longitudinal dataset concerning pig production. Analyses across the three case studies include a comparison with the original reports (for the two former datasets), as well as contrast with results from RF. The kernel framework not only allows a holistic view of data but also gives good results in each learning area. In unsupervised analyses, the main patterns detected in the original reports are conserved in kPCA. In supervised analyses SVM has better (or, in some cases, equivalent) performance than RF, while microbial signatures are consistent with the original studies and previous literature.

Objectives: The objectives of this study were to a) examine the distribution of chondrocalcinosis (CC), b) determine the risk factors of CC, and c) examine the radiographic phenotype of osteoarthritis (OA) associated with CC. Methods: Data from the Genetics of Osteoarthritis and Lifestyle (GOAL) study were used to describe the radiographic distribution of CC, and to conduct a case-control study in which cases with CC were compared with controls without CC. All participants had already completed a detailed questionnaire, been examined by a research metrologist, had radiographs of knees, hands, and pelvis, and had given urine and blood samples. All radiographs had been scored for structural radiographic changes of OA, and for the presence of CC. Frontal plane knee alignment was measured on all knee radiographs. The prevalence (95% confidence interval (CI)) of CC was calculated. The odds ratio (OR) and 95% CI were calculated for risk factors of CC, and for structural changes associated with CC in joints with OA. This was adjusted for age, gender, body mass index (BMI), and OA as appropriate, using logistic regression. Results: 3170 participants were included in this study. There were 431 cases with CC. The overall prevalence (95%CI) of CC in the GOAL population was 13.7% (12.5% - 14.9%). In the GOAL population, knee was the commonest site of CC. However, 42% of participants with CC did not have any knee involvement. There was evidence for a generalized predisposition to CC. For example, CC at one joint associated with CC at distant joints. Joints with CC clustered together more than would be expected by chance alone. At knees, wrists and hips, bilateral CC was more likely to associate with CC at distant joints than unilateral CC – also supporting the existence of a systemic predisposition to CC. After adjusting for confounding factors, there was an association between CC and increasing age, lower current BMI, and OA. The association between OA at one joint and CC at the same joint was present for all joints except for the hip. There was no association between CC and gender, diuretic intake, and selected single nucleotide polymorphisms in enzymes involved in pyrophosphate (PPi) metabolism. CC associated with peri-articular calcification, vascular calcification, low cortical bone mineral density (BMD) but not with low cancellous BMD. Self-reported arthroscopy, meniscectomy, knee injury, occupational knee joint loading and knee mal-alignment in the 3rd decade of life associated with knee CC. However, after adjusting for confounding factors including OA, there was no association between either self-reported or radiographically assessed current knee mal-alignment and knee CC. In joints with OA, the additional presence of CC at the same joint associated with a different radiographic phenotype of structural arthropathy. For example, in knees with OA, knee CC associated with attrition. In hips with OA, hip CC associated negatively with osteophytes, joint space narrowing, and sclerosis at the right hip but not at the left. Similarly, in wrists with OA, wrist CC associated with sclerosis in the right but not in the left wrist; in scapho-trapezioid joints (STJs) with OA wrist CC associated with sclerosis on both sides; in metacarpophalangeal joints with OA, wrist CC associated with cysts in the right but not in the left hand; and in 1st carpometacarpal joint with OA, wrist CC associated with cysts in the left but not in the right hand. In knees with OA, the additional presence of CC at distant joints associated with knee attrition. Those with knee CC + OA were excluded from this analysis to remove any local effects of CC. CC at distant joints did not associate with a distinct structural OA phenotype in other joints examined. Conclusion: These findings suggest that CC results form a systemic predisposition, and that it commonly occurs at other joints in the absence of knee involvement. Established risk factors of CC such as age, OA, and previous arthroscopy and/or meniscectomy were validated in this study. Several novel risk factors of CC e.g. low current BMI, low cortical BMD, and vascular calcification were identified. Several novel associations of knee CC i.e. early life knee malalignment, self-reported knee injury, and occupational knee loading were also recognised. There was convincing evidence to suggest that in joints with OA, the additional presence of CC modifies the OA phenotype, and that this varies from joint to joint.

This work examines the suitability of two different ontology approaches for the annotation of Schizosaccharomyces pombe (fission yeast) phenotypes derived from a number of screens of two fission yeast strain libraries- a temperature-sensitive library and an insertional library.

Pseudomonas tolaasii is the causal agent of brown blotch disease of mushrooms. The bacterium exists in two stable phenotypic forms: the smooth, virulent wild type (1116S), which produces the toxin tolaasin; and the rough, avirulent variant (1116R), which does not produce tolaasin. The PheN-PheS two-component signal transduction pathway regulates the phenotype of these forms. Spontaneous phenotypic switching between the 1116S and 1116R is mediated by a reversible 661 bp duplication within the sensor region of the pheN gene. The role of the environment in phenotypic switching between 1116S and 1116R was studied. It was found that frequency of switching from 1116R to 1116S in colonies was enhanced with the addition of mushroom extract to the medium. Characterisation of the active component of mushroom demonstrated the signal to be heat labile, active at 1,000 fold dilution and <12,000 kDa in size. Further characterisation suggested the presence of 2 distinct signal molecules in the mushroom extract, one in the cationic fraction and one in an anionic fraction. The toxin, tolaasin, a major determinant of phenotype is produced in a cell density dependent manner. The role of N-acyl homoserine lactone (AHL) mediated quorum sensing regulation of phenotype was studied. Thin layer chromatography indicated that 1116S produces the autoinducer compound, N -(3-oxododecanoyl)-L-homoserine lactone (OdDHL), while 1116R does not produce a detectable level of AHLs, suggesting that AHL production is under the regulation of the PheN-PheS two-component signal transduction pathway. Further investigation demonstrated the presence of an unidentified secondary quorum sensing autoinducer, which is only present below an OD₆₀₀ of 1.0. The addition of synthetic OdDHL to culture medium prior to inoculation with 1116S did not induce tolaasin synthesis, suggesting that either OdDHL is not involved with the regulation of tolaasin, or that a secondary level of regulation also exists.

Amnestic MCI (aMCI) is defined as a transitional stage in the development of dementia proper, with the most likely outcome being AD. This thesis endeavours to contribute to further clarification of the clinical signature of aMCI in a series of neuropsychological investigations based on the central memory disturbance. In the first stage of investigation, the nature of the subjective memory impairment was explored. An analysis based on a comprehensive questionnaire showed that memory complaints are common in clinical practice. The nature of these complaints showed that aMCI patients could not be separated from a ‘worried well’ group, although semantic dementia patients exhibited a clearly different profile. This highlights the necessity of objective assessment to corroborate memory complaints. The second stage of investigation explored the criterion for objective episodic memory impairment in aMCI. Addressing the purity of this deficit, tasks of object, people and face knowledge were used to probe for semantic memory impairment. The results showed that aMCI patients showed robust deficits compared to controls in these tests. Next, we looked into whether the episodic memory impairment could be reliably and robustly detected by probing associative learning memory. A key aspect of episodic memories is the rapid association of fragments into a coherent memory trace, which demands competence in associative learning. Based on work initiated in Cambridge, the nature of visuospatial associative learning impairment in aMCI was explored comparing the binding of novel and familiar stimuli. Recall was lower for the novel stimuli than for the familiar indicating a more sensitive task, since each element was new and required more encoding. Further, the prognostic utility of associative learning in combination with a measure of global cognition showed high sensitivity and specificity in predicting rapid progression to AD. Finally, we explored the utility of the associative learning deficit for differential diagnosis of early dementia. Differentiation of behavioural variant frontotemporal dementia (bvFTD) and AD has proved particularly difficult because of clinically overlapping symptoms. Compared to aMCI, bvFTD patients performed better overall, but inconsistently, on associative learning tasks suggesting that some bvFTD patients have subtle memory deficits. Still, associative learning may be a useful tool by which to explore differential diagnosis.

Meningioma and glioma are the most common primary brain tumors, but their etiologies are largely unknown. Although meningioma is usually benign, their intracranial location can lead to lethal consequences, and despite progress in surgery, radiotherapy, and chemotherapy the prognosis for patients with glioma remains poor. The only well-established environmental risk factor for meningioma and glioma is ionizing radiation. Evidence for inherited predisposition to meningioma and glioma is provided by a number of rare inherited syndromes where collectively these diseases account for only a small proportion of the twofold increased risk of brain tumors seen in first-degree relatives for meningioma and glioma patients. It is very possible that much of the excess familial risk is a consequence of co-inheritance of multiple low-risk genetic variations. With this in mind, the aims of the studies in this thesis were to discover genetic risk variants influencing the probability of acquiring the disease and to identify the association between risk variants on the tumor phenotype. To identify genetic variants influencing meningioma risk, a comprehensive tagging of the selected genes in a case-control study was performed. We identified nine risk variants in EGF, ERBB2, and LRIG2 genes. However, these findings could not be confirmed in another larger independent dataset. In addition, the study identified a correlation between LRIG2 protein expression and ER status when analyzed with different parameters. In a separate study with a larger sample of meningioma patients, the same correlation between LRIG2 and ER status was observed. To explore the potential association between reported germline risk variants and somatic genetic events, matched tumor and blood samples from glioma patients were analyzed by SNP array. The results identified correlations between EGFR gene variants and somatic aberrations within the EGFR locus and CDKN2A/B locus. To further study the relationship between germline risk variants and tumor phenotype, the same patient material was used and analyzed by three different techniques: SNP array, IHC, and FISH. The results revealed EGFR risk variants effecting copy number variation of the EGFR gene and the expression of the IDH1 and p53. Further comparison between different techniques such as SNP array and FISH analysis revealed the difficulty in achieving consistent results with different techniques. To summarize, the glioma studies show a link between genotype and phenotype where genetic risk variants in the EGFR gene were found to be associated with specific somatic aberrations. These associations are biologically interesting because EGFR is involved in multiple cellular processes. Additional studies of the direct functional role of these observations need to be conducted to elucidate the molecular mechanisms underlying the identified association between germline gene variants and somatic aberrations. For the meningioma studies, no significant risk variants influencing the disease were found but a correlation between LRIG2 and ER status was observed. This result suggests a potential role for the LRIG protein in the pathogenesis of meningioma, but more studies are needed to confirm this hypothesizes.

Cancer research foundation in northern Sweden and Lions cancer research foundation at Umeå university

The effects of androgens reach far and wide and can be physiological as well as pathological. They are not limited to males and involve almost every system in the human body. Their influence on reproductive development and behaviours is well studied, but more recently, attention has turned to the wider reaching consequences of androgen exposure. Disorders of sex development (DSD) are rare conditions in which individuals may be deficient in, or resistant to, the effects of androgens. The long-term health and quality of life for these individuals is not well reported, but where there are reports, there are descriptions of increased depressive like behaviours, anxiety and poor social functioning. Lack of androgens has been linked to poorer neurocognitive outcomes in some studies and there is a concern that more aggressive hormone replacement should be considered in early life for those individuals lacking in androgens. These disorders can be difficult to study for many reasons. Firstly, they are rare conditions. Secondly, adults with DSD do not tend to visit hospital regularly and can therefore be challenging to engage in research. Thirdly, studying the effects of early life exposure to steroid hormones and relating these to later life behaviours is incredibly complex. Animal models have been used for many years to study the hormonal environment. For my first study, I used a model of rodent neonatal androgen blockade by treating pups with the anti-androgen flutamide for the first five days of life. The animals were studied again in adolescence (6 weeks of age) and early adulthood (10 weeks of age). There were no significant differences found in testosterone, dihydrotestosterone and androstenedione levels in either age group, demonstrating that the androgen blockade was transient. The anogenital index (AGI) was significantly shorter in the treated animals when compared to controls at 6 weeks of age and 10 weeks of age. Phallus length was significantly shorter in treated males when compared to the healthy males at 6 weeks of age and at 10 weeks of age. Phallus weight was significantly lower in the treated animals at 10 weeks of age when compared to the healthy animals. This work demonstrated that my rodent model of neonatal androgen blockade was an effective one. My next study used the same rodent model and aimed to link the perinatal hormonal environment with in vivo brain chemistry using a painless, non-invasive technique known as Magnetic Resonance Spectroscopy. Using a mixed effects model, I analysed the effects of sex, gender, treatment with flutamide and age on the metabolite pattern of the rodent brain. Ɣ-aminobutyric acid (GABA), glucose, glutamine, glutamate, phosphocholine and myo-inositol all changed over time. The combined peaks of glutamate and glutamine also demonstrated a significant change over time. GABA, glutamate, phosphocholine and myo-inositol showed significant sex differences as did the combined peaks of glycerophosphocholine and phosphocholine, N-acetylaspartate (NAA) and N-acetylaspartylglutamate (NAAG) and glutamate and glutamine. Aspartate, GABA and myo-inositol were all significantly changed by treatment of the animals with flutamide and GABA and myo-inositol levels in treated males were similar to control females at both 6 and 10 weeks. My final study using the rodent model of androgen blockade looked at the histological changes in the brain. Brains were sectioned and stained for neuronal cell counts and microglial cell counts, and PCR for the Androgen Receptor (AR) was performed. I demonstrated significant, sexually dimorphic changes in neuronal cell counts, microglial cell counts and androgen receptor expression in two clearly defined areas in the rodent brain. In summary, my rodent work demonstrated a link between the neonatal hormonal environment and the sexually dimorphic chemistry and histology of the in vivo brain, and supports the hypothesis that the microglial cell plays a critical role in brain masculinisation. To include a translational aspect to this thesis I extended my work to a population of undermasculinised boys, who were attending hospital for an hCG stimulation test as part of their investigations for 46 XY DSD. The hCG stimulation test is a valuable method for assessing androgen production but there is a need to explore its utility in assessing androgen responsiveness and long-term prognosis. I aimed to assess the effects of the hCG test on the in vivo brain chemistry using MRS, and the peripheral transcriptome using microarray. I reliably demonstrated metabolites in the brains of healthy male infants, healthy female infants and affected male infants. Healthy male infants had significantly lower levels of N-acetylaspartate than affected males in the hypothalamus and lower levels of the phosphocholines in the frontal cortex. In my transcriptomic study of DSD patients, I demonstrated the existence of an androgen responsive group of small RNAs that are measurable in peripheral mononuclear blood cells, and that change over the short duration of an hCG stimulation test, raising the prospect of combining the biochemical assessment of testosterone production with an objective molecular assessment of androgen sufficiency. In summary, in this thesis I have successfully linked the early hormonal environment with later life in vivo brain chemistry, confirmed by histological studies. I have also identified a novel marker, which could potentially be used as an assessment of androgen sufficiency in the future.

Psychiatric symptoms are more prevalent in Huntington's disease (HD) than the general population, but reasons for this are unknown. The primary aim of this research was to investigate possible familial influences on the psychiatric phenotype in HD. 96 gene positive and 5 gene negative siblings were recruited from 50 HD families throughout the UK and underwent a lifetime psychiatric history assessment using semi-structured interview and case-note review. Gene positive index individuals had high lifetime rates of depressive (56%) and anxiety (38%) disorders. Their depressive episodes were less severe and more frequent with an older age of onset and fewer biological symptoms than individuals with depression without HD. Within gene positive sibling-pairs (n=53), there was significant familial aggregation of the presence (ĸ=0.46, \(p\)=0.004) and course (ICC=0.47, \(p\)=0.002) of depressive disorders and the presence of irritability (ĸ=0.357, \(p\)=0.024) and aggression (ĸ=OJ84, \(p\)=O.Ol6). Two gene negative siblings had lifetime psychiatric diagnoses. The high prevalence of psychiatric co-morbidity in HD cannot be entirely explained by the HD gene. Familial factors, most likely other genetic factors, are likely to play a role. Further research into the contribution of biological and environmental factors to the psychiatric phenotype in large samples of individuals with HD is warranted.

This thesis explores genotype-phenotype relations in lupus nephritis. Over the past decade, advances in genetic research have improved our understanding of the genes that contribute to lupus susceptibility. However a clear understanding of how these genetic factors interact with the immune system to produce disease in a given patient or group of patients with SLE is difficult to define at this time. Recent monogenic causes of SLE have provided valuable insights into lupus pathogenesis. A comprehensive clinical characterisation of 164 biopsy proven lupus nephritis patients was undertaken in the initial phase of this research. 27% of patients recruited were of juvenile onset, as defined by diagnosis of nephritis before 18 years of age. Non-European lupus nephritis patients had a younger age of onset and higher rate of progression to end-stage renal disease. Sixteen individuals had a first degree family history of lupus nephritis including five sibling pairs. Familial clustering of nephritis was associated with juvenile onset disease and a more severe clinical phenotype. The entire cohort was genotyped by ImmunoChip for known lupus susceptibility polymorphisms. Associations with lupus nephritis were found in polymorphisms in the HLA region, IRF5, ITGAM, STAT4, TNFAIP3, TNFSF4 and ETS1. Whole Exome Sequencing of familial lupus nephritis cases identified a large number of potential candidates for functional investigation. A promising short list of candidates was created using pedigree information, focusing on variants predicted to be damaging and by incorporating prior knowledge of the biologic pathways implicated in lupus. Candidates under consideration for functional testing include genes involved in activation of Ras pathways, genes involved in the antioxidant defence system, variants in ubiquitination pathways, mutations in serine/threonine protein kinases and genes involved in toll-like receptor and type I interferon signalling pathways.

Among the most fundamental features shared by all organisms is the mapping of information encoded in genotypes (genetic material) to generate phenotypes (biological structures, functions, and traits). Hence, elucidating the structure of genotype-phenotype (GP) maps is important for understanding evolution and biology. While it is known that often GP maps are highly degenerate with many different genotypes adopting the same phenotype, the distribution of genotypes over phenotypes is less well studied. In this thesis we investigate the question of the distribution of genotypes over phenotypes, or put differently the distribution of neutral set sizes (NSS), where a neutral set is the collection of all genotypes in a GP map which map to the same phenotype. We focus on examining phenotypic bias in GP maps, where some phenotypes have disproportionally large NSS as compared to others. We find phenotypic bias to be ubiquitous in the broad range of GP maps that we analyse, from the genetic code up to molecular RNA to a model of neuronal connections, and hence we hypothesise bias to be a common property of GP maps. Further, we also consider the implications that this bias has for evolutionary outcomes, and we argue that bias is a significant influencing factor in determining evolutionary outcomes. Finally, we propose a method to predict a phenotype's NNS via estimating the phenotype's structural complexity, without using detailed knowledge about the specifics of the relevant GP map. We achieve this via a novel application of algorithmic information theory and especially Levin's coding theorem.

Objective: The human immunodeficiency virus type 1 (HIV-1) currently infects over 30 million people worldwide. It uses one of two main co-receptors to infect cells. The primary objective of this thesis is to evaluate genotypic and phenotypic assays for co-receptor usage in the clinical setting and investigate approaches for improvement of these assays. Methods: The concordance of recombinant co-receptor phenotyping assays and the predictive ability of genotype-based methods including the ‘11/25’ rule, position specific scoring matrices (PSSMs), and support vector machines (SVMs) were evaluated in the clinical setting using patient-derived plasma samples. Samples and patient data were evaluated in cross-sectional analyses from a retrospective population-based cohort of HIV-infected individuals enrolled in the HIV/AIDS Drug Treatment Program in British Columbia, Canada. Results: Current implementations of HIV V3 region-based predictors for HIV co-receptor usage tested on patient derived samples are inadequate in the clinical setting, primarily due to low sensitivities as a result of difficult to detect minority species. Recombinant phenotype assays also show discordances when tested against each other on the same set of patient derived samples, raising doubts if any of these assays can truly be considered a ‘gold standard’. Significant associations between clinical progression, viral sequence-based predictors of co-receptor usage and the output of recombinant assays are observed, suggesting that sensitivity can be improved by incorporating CD4% into genotype-based predictors. This is verified with a SVM model which showed a 17% increase in sensitivity when CD4% was incorporated into training and testing. Conclusion: This work in this thesis has exposed the difficulty in determining the co-receptor phenotype in the clinical setting, primarily due to minority species. Although genotypic methods of screening for HIV co-receptor usage prior to the administration of CCR5 antagonists may reduce costs and increase turn-around time over phenotypic methods, they are currently inadequate for use in the clinical setting due to low sensitivities. Although the addition of clinical parameters such as CD4 count significantly increases the predictive ability of genotypic methods, the presence of low-levels of X4 virus continues to reduce the sensitivity of both genotypic and phenotypic methods.

El gen Fas apoptotic inhibitory molecule 1 (FAIM1) fue descubierto en las células B del sistema y está evolutivamente conservado. FAIM1 codifica al menos para cuatro isoformas, de entre ellas FAIM-S y FAIM-L son las más relevantes. FAIM-L está codificada por splicing alternativo, se expresa sólo en neuronas y presenta veintidós aminoácidos más que la isoforma descrita en células B, FAIM-S. Ambas proteínas fueron identificadas como antagonistas de los receptores de muerte en diferentes células, siendo FAIM-L la única isoforma capaz de ejecutar una acción protectora en neuronas. Recientemente se ha revelado que las isoformas de FAIM1 también protegen contra la muerte inducida por estrés oxidativo. Estás proteínas ejercen adicionalmente otras funciones no relacionadas con la apoptosis. FAIM-S regula distintas funciones en el sistema inmune, participa en el metabolismo lipídico y glucolítico, en la agregación proteica y en el crecimiento neurítico. FAIM-L modula la transmisión y la plasticidad sináptica y la neurodegeneración axonal. Además, las isoformas de FAIM1 son importantes en diferentes patologías como el mieloma múltiple, Alzheimer y la obesidad. Aunque diversas funciones de FAIM1 han sido definidas en el sistema nervioso, todavía quedan muchas preguntas por elucidar sobre su papel en el sistema nervioso. Por ello, decidimos estudiar el fenotipo neuronal del ratón FAIM-KO. Inesperadamente observamos que los ratones FAIM-KO desarrollaban convulsiones tras ser expuestos a estímulos sensoriales y/o estresantes y que esta predisposición se iniciaba en la edad adulta. Este hallazgo nos llevó a estudiar qué alteraciones podían favorecer el desarrollo de estas crisis convulsivas. Observamos que los ratones FAIM-KO que convulsionaban presentaban alteraciones en el hipocampo típicas de cerebro epilépticos tales como aumento de la neurogénesis, expresión ectópica del neuropéptido Y y aumento de la expresión de c-fos en el hipocampo. Sin embargo, estos cambios no aparecían en los ratones sin convulsiones. Aunque los ratones FAIM-KO no presentaban inflamación o muerte celular, su densidad de células gliales estaba ligeramente disminuida y la densidad de ciertas poblaciones de interneuronas también estaban alteradas en el hipocampo. Además, la expresión de algunas proteínas sinápticas (SNAP25 y vGLUT) y apoptóticas (Fas y XIAP) así como en el número de dendritas primarias de las neuronas granulares estaban alterados en el hipocampo de estos ratones. En cuanto a su comportamiento, estos ratones eran más activos y tenían una mayor predisposición a interaccionar con otros ratones, mientras que presentaban deficiencias en sus habilidades cognitivas y de construcción del nido. Estos resultados revelan interesantes nuevas funciones de FAIM1 en el cerebro, sin embargo, es necesario corroborar el papel de esta proteína en el desarrollo de convulsiones ya que los ratones FAIM-KO usados en este estudio tienen un fondo genético mixto.
Fas apoptotic inhibitory molecule 1 (FAIM1) gene is highly conserved in the evolution and was firstly characterized in B cells in 1999. FAIM1 codifies for at least four isoforms, among which FAIM-L and FAIM-S are the most relevant. FAIM-L is formed by alternative splicing, is exclusively expressed in neurons and contains twenty-two amino acids more than the first isoform characterized in B cells, FAIM-S. The first described role of FAIM1 was this protective effect against death-receptor mediated cell death. FAIM-S and FAIM-L exert this function in different cells, and FAIM-L is the only FAIM1 isoform that protects neurons against induced apoptosis. Recently, a protective role of FAIM1 in cellular stress-induced cell death has been reported. FAIM1 have been also implicated in non-apoptotic functions. FAIM-S participates in different non-apoptotic functions in the immune system, in glucose and lipid metabolism, in protein aggregation and in neurite outgrowth. Otherwise, FAIM-L acts as a regulator in synaptic transmission, axonal degeneration and synaptic plasticity process of long-term depression. Besides, FAIM1 isoforms have relevant roles in different diseases. In that sense, FAIM-S is deregulated in multiple myeloma and obesity and FAIM-L is downregulated in Alzheimer’s disease. Although some FAIM1 isoforms functions have been reported in nervous system, many questions about their actions in CNS are still a high exciting mystery. Therefore, we decided to characterize the neuronal phenotype of FAIM-KO mice to unravel FAIM1 functions in brain. Surprisingly, we observed age-dependent sensory-induced seizures in FAIM-KO mice. Owing to that finding, the work was focused to unravel mechanisms related with seizure susceptibility in FAIM-KO mice. FAIM-KO mice with seizure showed typical hippocampal cellular and molecular alterations reported in epilepsy models such ectopic neuropeptide Y expression, increase neurogenesis and increase c-fos expression in hippocampus. However, these effects have not been observed in non-convulsive FAIM-KO mice. Although, neuroinflammation and cell death were not apparent in FAIM-KO mice, these animals exhibit a decrease in glial density in hippocampus and alterations in parvalbumin- and calretinin-containing interneuron populations. These mice also exhibited slightly changes in the expression of synaptic proteins SNAP25 and vGLUT1, deregulation in mRNA levels of Fas and XIAP, and alteration in primary dendrites of granule cells in hippocampus. Interestingly, FAIM-KO mice were hyperactive, had impairment in cognitive tasks and in nest construction and exhibited an increase in social interactions. These results point to new exciting roles of FAIM1 in CNS. However, the use in this work of FAIM-KO mice derived of mixed background makes the replication of these experiments in other genetic background necessary for ensuring a role of FAIM1 in seizure susceptibility.

Development of the aortic arch arteries from the pharyngeal arch arteries is a complex process requiring the interaction of several tissue types and tightly regulated gene expression during embryogenesis. In this work, the role of transcription factor Pax9 in cardiovascular development was investigated using transgenic mice, imaging techniques and gene expression analysis. Pax9 is specifically expressed in the pharyngeal endoderm at mid-embryogenesis, and Pax9-null mice die at birth from defects in the formation and remodeling of the 3rd and 4th pharyngeal arch arteries resulting in absent common carotid arteries, aberrant right subclavian artery and interrupted aortic arch.

Stage II/III CRCs are currently treated by surgery, then stage III and stage II patients exhibiting high-risk pathological features undergo a course of 5-FUbased adjuvant chemotherapy, often with oxaliplatin added. Improved characterisation of stage 11/111 tumours is required in order to improve their treatment and to ensure that drugs are only given to patients where there is evidence of a likely response. This will reduce the unnecessary and costly overtreatment of these patients, as well as the associated unpleasant side effects. Pyrosequencing, immunohistochemistry, FISH, and next-generation sequencing were used in order to assess a number of molecular markers of high-risk stage 11 CRCs in two cohorts. Of the markers tested, only mutation in KRAS was prognostic in stage 11/11I CRC (HR = 1.4). Mutation of BRAF, NRAS, PIK3CA and overexpression/amplification of HER-2 were not prognostic or predictive of response to 5-FU chemotherapy in the QUASAR-1 trial. TOP2A was coamplified with HER-2 in 28% cases. Copy number profiles produced from this cohort showed the differences between dMMR and pMMR, and the effect of high stromal content on CNV. Cases from the QUASAR-1 fit into one of three groups regarding CNV which need further characterisation and their relationship to prognosis needs to be established. KRAS mutation is a prognostic marker in stage 11/11I disease and could be tested for routinely. Importantly none of the other markers tested predicted a worse outcome with 5-FU-based chemotherapy. TOP2A is frequently coamplified with HER-2 suggesting that a subset of these cases may respond to anthracycline therapy.

Development of the aortic arch arteries from the pharyngeal arch arteries is a complex process requiring the interaction of several tissue types and tightly regulated gene expression during embryogenesis. In this work, the role of transcription factor Pax9 in cardiovascular development was investigated using transgenic mice, imaging techniques and gene expression analysis. Pax9 is specifically expressed in the pharyngeal endoderm at mid-embryogenesis, and Pax9-null mice die at birth from defects in the formation and remodeling of the 3rd and 4th pharyngeal arch arteries resulting in absent common carotid arteries, aberrant right subclavian artery and interrupted aortic arch. The aim of the work presented in this thesis was to identify genes that potentially acting in the same genetic regulatory network as Pax9 to control cardiovascular development. The transcription factor Msx1 was selected as this was shown to interact with Pax9 in tooth development, and transgenic mice lacking Msx1 and Msx2 display cardiovascular defects. Surprisingly, this analysis shows that mice simultaneously null for Pax9 and heterozygous for Msx1 have a significantly reduced incidence of aortic arch defects compared to Pax9 –/– mice. Although Pax9 and Msx1 proteins can interact during tooth formation, the expression of these two genes does not overlap in the pharyngeal arches, with Msx1 being expressed in neural crest cells. Immunohistochemistry labelling revealed that more neural crest cells were present within the 3rd pharyngeal arch of Pax9 –/–;Msx1+/– embryos compared to Pax9 –/– embryos, which suggesting a role for these cells in protecting the 3rd pharyngeal arch artery from regressing as observed in Pax9 –/– embryos. In addition, a thick layer of smooth muscle around the 3rd PAA in Pax9 –/–;Msx1+/– embryos compared to Pax9 –/– embryos was observed, , all together indicating that neural crest cells differentiate to smooth muscle in Pax9 –/–;Msx1+/– embryos and this help protect and support the 3rd PAA during remodeling. Our data therefore show that Msx1 heterozygosity appears to modulate the Pax9-null cardiovascular phenotype by a yet unrecognized mechanism.

Early growth response protein-1 (EGR-1) is a transcription factor induced by stress or injury, mitogens, and differentiation factors. It has been shown to be regulated by various cytokines, growth factors and by ischemic/hypoxic stress as well as shear stress and mechanical injury. These regulators have been linked to both the development as well the degeneration of the musculoskeletal system, namely articlar cartilage, intervertebral discs (IVDs) and bone. Furthermore, Egr-1 has been shown to regulate the expression of collagens and enzymes affecting the extracellular matrix. Polymorphisms of DNA binding sites for EGR-1 have shown to be associated with both disc degeneration and osteoporosis. The aim of this study was to determine the affects of EGR-1 deficiency on articular cartilage, IVD and bone phenotype.
Wild-type (+/+) C57Bl/6 or Egr-1-deficient (−/−) mice were sacrificed at the same age interval (8- to 9-months). Standard histological preparation and staining with Safranin-O/Fast Green were done. Also immuncohistochemistry was performed using anti-bodies to type X collagen, cleavage products of both type II collagen and aggrecan. Imaging of mice was with plain radiographs, bone mineral density measurements and microCT analysis.
Results revealed that these mice have differences including abnormal bone structure and density, structural and possibly compositional differences in articular cartilage and structural and biochemical changes in IVDs. This points to the importance of Egr-1 in the maintenance of normal bone, IVD and articular cartilage and makes it a possible target for initiating pathological conditions of these tissues.
La protéine de croissance EGR-1 (Early Growth Response protein-1, en anglais) est un facteur de transcription qui est induit par la tension ou la blessure, les facteurs mitogènes, et les facteurs de différenciation. EGR-1 est ainsi régulé par divers cytokines, facteurs de croissance, par les conditions ischémique, ainsi que la tension et les blessures mécaniques. Ces régulateurs ont été reliés au développement ainsi que la dégradation du système squeletto-musculaire, particulièrement le cartilage articulaire, les disques intervertébraux (DIV) et l'os. De plus, il a été démontré que EGR-1 peut réguler l'expression des collagènes et d'enzymes contribuant à la matrice extracellulaire. Le polymorphisme de séquences d'ADN pour les des sites d'attachements d'EGR-1 a démontré être associé avec la dégradation de disques intervertébraux et l'ostéoporose. L'objectif de cette étude était de déterminer l'effet d'une expression réduite d'EGR-1 sur les phénotypes du cartilage articulaire, les DIV, et l'os.
Les souris C57Bl/6 de phénotype sauvage (+/+) ou ceux avec une expression réduite d'EGR-1 (−/−) ont été sacrifiées au même intervalle d'âge (8 à 9 mois). La préparation histologique standard et la coloration avec Safranin-O/Fast Green a été fait. Aussi l'immunohistochimie a été exécuté avec des anticorps pour le collagène de type X, et les produits de clivage du collagène de type II ainsi que les aggrécanes. L'imagerie de souris a été faite avec les radiographies simples, les mesures de densité minéraux de l'os, et avec l'analyse de micro-tomodensitomètre.
Les résultats ont révélé que ces souris ont des différences incluant la structure et densité d'os anormaux, les différences structurelles et possiblement compositionnelles dans le cartilage articulaire, et les changements structurels et biochimiques dans les DIV. Ceci indique à l'importance d'EGR-1 dans l'entretien d'os normal, des DIV et le cartilage articulaire, et le rend une cible possible pour initier les conditions pathologiques de ces tissus.

The phenotype inference from genotype in RNA viruses maps the viral genome/protein sequences to the molecular functions in order to understand the underlying molecular mechanisms that are responsible for the function changes. The inference is currently done through a laborious experimental process which is arguably inefficient, incomplete, and unreliable. The wealth of RNA virus sequence data in the presence of different phenotypes promotes the rise of computational approaches to aid the inference. Key residue identification and genotype-phenotype mapping function learning are two approaches to identify the critical positions out of hitchhikers and elucidate the relations among them. The existing computational approaches in this area focus on prediction accuracy, yet a number of fundamental problems have not been considered: the scalability of the data, the capability to suggest informative biological experiments, and the interpretability of the inferences. A common scenario of inference done by biologists with mutagenesis experiments usually involves a small number of available sequences, which is very likely to be inadequate for the inference in most setups. Accordingly biologists desire models that are capable of inferring from such limited data, and algorithms that are capable of suggesting new experiments when more data is needed. Another important but always been neglected property of the models is the interpretability of the mapping, since most existing models behave as ’black boxes’. To address these issues, in the thesis I design a supervised combinatorial filtering algorithm that systematically and efficiently infers the correct set of key residue positions from available labeled data. For cases where more data is needed to fully converge to an answer, I introduce an active learning algorithm to help choose the most informative experiment from a set of unlabeled candidate strains or mutagenesis experiments to minimize the expected total laboratory time or financial cost. I also propose Disjunctive Normal Form (DNF) as an appropriate assumption over the hypothesis space to learn interpretable genotype-phenotype functions. The challenges of these approaches are the computational efficiency due to the combinatorial nature of our algorithms. The solution is to explore biological plausible assumptions to constrain the solution space and efficiently find the optimal solutions under the assumptions. The algorithms were validated in two ways: 1) prediction quality in a cross-validation manner, and 2) consistency with the domain experts’ conclusions. The algorithms also suggested new discoveries that have not been discussed yet. I applied these approaches to a variety of RNA virus datasets covering the majority of interesting RNA phenotypes, including drug resistance, Antigenicity shift, Antibody neutralization and so on to demonstrate the prediction power, and suggest new discoveries of Influenza drug resistance and Antigenicity. I also prove the extension of the approaches in the area of severe acute community disease.