Over half the human genome consists of repetitive sequences. One major class is the tandem repeats (TRs), which are defined by their location in the genome, repeat unit, and copy number. TRs loci that exhibit variant copy numbers are called Variable Number Tandem Repeats (VNTRs). High VNTR mutation rates of approximately 0.0001 per generation make them suitable for forensic studies, and of interest for potential roles in gene regulation and disease. TRs are generally divided into three classes: 1) microsatellites or short tandem repeats (STRs) with patterns <7 bp; 2) minisatellites with patterns of seven to hundreds of base pairs; and 3) macrosatellites with patterns of >100 bp. To date, mini- and macrosatellites have been poorly characterized, mainly due to a lack of computational tools. In this thesis, I utilize a tool, VNTRseek, to identify human minisatellite VNTRs using short-read sequencing data from nearly 2,800 individuals and developed a new computational tool, MaSUD, to identify human macrosatellite VNTRs using data from 2,504 individuals. MaSUD is the first high-throughput tool to genotype macrosatellites using short reads. I identified over 35,000 minisatellite VNTRs and over 4,000 macrosatellite VNTRs, most previously unknown. A small subset in each VNTR class was validated experimentally and in silico. The detected VNTRs were further studied for their effects on gene expression, ability to distinguish human populations, and functional enrichment. Unlike STRs, mini- and macrosatellite VNTRs are enriched in regions with functional importance, e.g., introns, promoters, and transcription factor binding sites. A study of VNTRs across 26 populations shows that minisatellite VNTR genotypes can be used to predict super-populations with >90% accuracy. In addition, genotypes for 195 minisatellite VNTRs and 22 macrosatellite VNTRs were shown to be associated with differential expression in nearby genes (eQTLs). Finally, I developed a computational tool, mlZ, to infer undetected VNTR alleles and to detect false positive predictions. mlZ is applicable to other tools that use read support for predicting short variants. Overall, these studies provide the most comprehensive analysis of mini- and macrosatellites in human populations and will facilitate the application of VNTRs for clinical purposes.

>Magister Scientiae - MSc / The generation of biomedical research data on the African continent is grow- ing, with numerous studies realizing the importance of African genetic diver- sity in discoveries of human origins and disease susceptibility. The decrease in costs to purchase and utilize such tools has enabled research groups to produce datasets of signi cant scienti c value. However, this success story has resulted in a new challenge for African Researchers and institutions. An increase in data scale and complexity has led to an imbalance of infrastructure and skills to manage, store and analyse this data. The lack of physical infrastructure has left genomic research on the continent lagging behind its counterparts abroad, drastically limiting the sharing of data and posing challenges for researchers wishing to explore secondary analysis, study veri cation and amalgamation. The scope of this project entailed the design and implementation of a proto- type genome archive to support the e ective use of data resources amongst researchers. The prototype consists of a web interface and storage backend for users to upload and browse projects, datasets and metadata stored in the archive. The server, middleware, database and server-side framework are components of the genome archive and form the software stack. The server component provides the shared resources such as network connectivity, le storage, security and metadata database. The database type implemented in storing the metadata relating to the sample les is a NoSQL database. This database is interfaced with the iRods middleware component which controls data being sent between the server, database and the Flask framework. The Flask framework which is based on the Python programming language, is the development platform of the archive web application. The Cognitive Walkthrough methodology was used to evaluate suitabil- ity of the software for its users. Results showed that the core conceptual model adopted by the prototype software is consistent and that actions available to the user are visible. Issues were raised pertaining to user feedback when per- forming tasks and metadata term meaning. The development of a continent wide genome archive for Africa is feasible by utilizing open source software and metadata standards to improve data discovery and reuse.

Indiana University-Purdue University Indianapolis (IUPUI) / The main focus of this study is to determine the impact of Set-Binding Factor 2 (SBF2) on human-derived neurons in the context of taxane-induced peripheral neuropathy. Taxane-induced peripheral neuropathy (TIPN) is a devastating survivorship issue for many cancer patients; SBF2 has been previously identified as a potential germline predictor that has been found to be significantly associated with severe TIPN in African American (AA) patients. The work described here provides ex vivo support for the use of SBF2 as a genotypic biomarker to identify a priori which patients are at a higher risk of manifesting severe TIPN. This study demonstrates that diminished expression of SBF2 exacerbated the effect of paclitaxel on viability and morphology and altered the functional response of a neuronal model exposed to paclitaxel treatment. Furthermore, transcriptomic work showed that reduced expression of SBF2 in a neuronal model treated with paclitaxel impacted the expression of genes that modulate stress-induced cell death and pain threshold. Altogether, these findings suggest that SBF2 plays a role in the development of TIPN. This work sheds light on the pathways potentially involving SBF2 that can be studied to further evaluate the function of this gene in neurons and its contribution to severe TIPN. Further functional approaches investigating these pathways will be pivotal in elucidating the underlying biological mechanism for this toxicity and identifying novel targeted therapeutic strategies to prevent or treat TIPN. / 2021-05-17

京都大学 / 0048 / 新制・課程博士 / 博士(医科学) / 甲第21023号 / 医科博第84号 / 新制||医科||6(附属図書館) / 京都大学大学院医学研究科医科学専攻 / (主査)教授 斎藤 通紀, 教授 萩原 正敏, 教授 小川 誠司 / 学位規則第4条第1項該当 / Doctor of Medical Science / Kyoto University / DFAM

No description available.

This thesis explores the genetic architecture of malting quality within the Virginia Tech barley breeding program, and discusses implications for imposing selection on complex traits that are difficult to phenotype. Malting quality measures are destructive, and can not be performed before selection must be made for advancement of breeding lines in winter barley. A growing body of evidence suggests that malt quality is influenced by malting regime, growing environment, line genotype, and the interactions between them. We aim to better understand the genetic effect on malt quality in two manners: first, as it relates to the genetic architecture regulating malt quality parameters, and second the relationship between genetic growth patterns to end-use malting traits. This study included two years of breeding trial data of two and six-row winter malt barley across two locations. Results of a genome-wide association scan and genomic prediction of malt quality traits indicated that they are largely quantitative traits with complex inheritance. Previous studies have identified quantitative trait loci and genes regulating malt quality traits in markedly different germplasm. Heritability of traits ranged from 0.27 to 0.72, while mean predictive abilities ranged from 0.45 to 0.74. Thus, selection on genomic estimated breeding values (gEBVs) should perform similarly to selection on single phenotypic observations of quality, but can be done within the same season. This indicates that genomic selection may be a viable method to accelerate genetic improvement of malting quality traits. The use of gEBVs requires that lines be genotyped with genome-wide markers, somewhat limiting the number of candidate individuals. Selection on growth and development traits genetically correlated with quality measures could allow for selection among a much greater number of candidates if high-throughput phenotypes can be collected on many ungenotyped indivduals. Growth and development was quantified by the near-infrared vegetation index (NDVI) extracted from aerial images captured from multiple time points throughout the growing season. Estimates of genetic correlation identified time points throughout the season when quality traits are related to growth and development. We demonstrated that aerial imagery can discern growth patterns in barley and suggest ways it can be incorporated into the breeding pipeline. / Master of Science / Malt barley (Hordeum vulgare) is the preferred source of fermentable sugar used to brew beer. Currently, the majority of malt barley used in the United States is grown in the upper mid-west or imported from Europe. The east coast could become a producing region if high quality, disease resistant varieties were available to growers. The Virginia Tech small grains breeding program began breeding locally adapted malt barley in 2010. This project aims to improve the breeding process by incorporating information from genomic sequencing, malt quality and aerial imagery. Malt barley differs from that used for animal feed or human food because specific quantities of starches, proteins, and enzymes are necessary in the brewing process. The quantity of these molecules are determined through lab analysis and determine the grain's suitability for particular brewing styles. This analysis is timeconsuming and costly because it involves a three-step process of malting the grain, brewing with the malt, and analyzing the wort. The wort is the liquid sugar solution which is produced by heating the malt with water to a high temperature in a process called 'mashing'. Lab quality analysis for the thousands of lines evaluated in a breeding program in any given year is unfeasible. However, by understanding the genetic regulation of malt quality traits, breeders can employ techniques like genomic selection to improve these traits in a shorter amount of time. Additionally, this work identifies relationships between growth and quality. The grain is the result of the plant's growth throughout the entirety of the season. Measuring growth repeatedly through time was previously difficult until the advent of aerial imagery. Images captured from drones have been used to quantify growth in a variety of plants, but is not extensively done in malt barley. Relating growth to quality will help breeders understand genetic patterns of growth and development which may be advantageous in the production of high quality malt barley.

Oxidative stress (OS) occurs when DNA repair mechanisms are overcome by the amount of single and double strand DNA breaks caused by an accumulation of reactive oxygen species (ROS). Genomic instability (GI) by microsatellite instability (MSI) accumulation is characterized by changes in DNA single tandem repeats (STR) as a direct result of ROS. Deregulation of DNA repair and tumor suppressor pathways have been described as causes of tumor progression and metastasis. Studies in mammals have focused on GI and the implications of increased mutation frequency due to accumulation of MSI leading to development of diseases, including infertility and cancer. Ovarian cancer is a deadly disease displaying the highest mortality rate among gynecological cancers. Hereditary ovarian cancer displays GI that can be established early in primordial germinal cells (PCGs) development and migration across the genital ridge, where PGCs are exposed to ROS damage. The hypothesis of this study was ROS-induced GI is marked by the accumulation of MSI on repetitive sequences of DNA that override DNA repair, tumor suppressor and redox homeostasis pathways. In this study, we induced ROS in human ovarian cell lines by hydrogen peroxide (H2O2) exposure, as well as evaluated mouse PGCs to determine whether MSI occurs in specific regions of human and mouse genomes. Our results show that MSI was present in specific markers after ROS-induced damage in human ovarian cells and in mouse Sod1 knockout PGCs during cell migration, both of which accumulate specific mutations caused by free radical damage. Ovarian tumor cells and mouse PGCs showed an increase of MSI in 12 human and 5 mouse repetitive markers that are located near important genes related to DNA repair, tumor suppression, cell proliferation, apoptosis and differentiation. This could be a signal that leads to tumor initiation, formation and progression in adult ovarian cells due to improper DNA repair and tumor suppression mechanisms or in disruption of PGC migration that determines germinal cell pool selection during early embryonic development due to absence of cell antioxidant mechanisms. Therefore, these specific unstable STRs are novel biomarkers that could be useful in early diagnostics, prognosis, and successful therapy of ovarian tumorigenesis.

No description available.

Deep learning models are dominating performance across a wide variety of tasks. From protein folding to computer vision to voice recognition, deep learning is changing the way we interact with data. The field of natural products, and more specifically genomic mining, has been slow to adapt to these new technological innovations. As we are in the midst of a data explosion, it is not for lack of training data. Instead, it is due to the lack of a blueprint demonstrating how to correctly integrate these models to maximise performance and inference. During my PhD, I showcase the use of large language models across a variety of data domains to improve common workflows in the field of natural product drug discovery. I improved natural product scaffold comparison by representing molecules as sentences. I developed a series of deep learning models to replace archaic technologies and create a more scalable genomic mining pipeline decreasing running times by 8X. I integrated deep learning-based genomic and enzymatic inference into legacy tooling to improve the quality of short-read assemblies. I also demonstrate how intelligent querying of multi-omic datasets can be used to facilitate the gene signature prediction of encoded microbial metabolites. The models and workflows I developed are wide in scope with the hopes of blueprinting how these industry standard tools can be applied across the entirety of natural product drug discovery. / Thesis / Doctor of Philosophy (PhD)

Biological evidence may contain any number of cells in any proportion. Extreme low-template DNA samples are often very difficult to interpret due to complex signal or peaks which may be indistinguishable from baseline noise. Current solutions focus on increasing the amount of amplicon detected by adjusting PCR cycle number or capillary electrophoresis injection parameters. Consensus profiling is an additional option. However, the aforementioned solutions are often not helpful for extreme low-template samples due to the high occurrence of allelic drop-out. Additionally, PCR is a destructive technique that causes one amplification to completely exhaust this type of sample, making further typing and analysis impossible. Therefore, a technique that allows for the re-generation of a DNA template in order to amplify it multiple times would be an extremely useful tool. This study outlines the development of a method that allows for the recursive amplification of a DNA sample. Amplification was performed using biotinylated primers for an STR locus and the resulting product was cleaned using streptavidin-coated magnetic beads to sequester the amplicons. Subsequent centrifugal filtration was used to remove the remaining PCR components, thus isolating the original genomic DNA. Re-amplification was then successfully performed at a different STR locus. Though successful, multiple run-throughs of the method indicated retention of signal from the original amplification as well as significant genomic DNA loss during the process. This study outlines experiments seeking to characterize the cause(s) of these imperfections in order to effectively direct method optimization. A computer generated dynamic model was also created and used to simulate the recursive amplification process to assist in development. When optimized, it is expected that recursive amplification can significantly reduce the difficulties associated with low-template DNA analysis and eradicate the concept of an ‘exhaustive’ DNA sample.