Texas Digital Library
UT Southwestern Medical Center Institutional Repository (University of Texas)Not a member yet
10274 research outputs found
Sort by
Progress in polycystic kidney disease: enhanced assessment & emerging therapies
Get PDFDetailed formal protocol with illustrations and extensive bibliography.A recording of the protocol presentation is available on UT Southwestern's Mediasite. Note: Access to the video is restricted to authorized UT Southwestern users only.UT Southwestern--Internal Medicin
Correcting Biases in Multi-Module Neural Networks, Through Efficient Hyperparameter Optimization and Statistically Meaningful Uncertainty Quantification, with Applications to Neurological Disorders
No full textDeep learning is a branch of machine learning that employs artificial neural networks to produce inferences from data. These networks have been successfully applied to a plethora of clinically and biologically related problems, including prognosis and diagnosis for a broad spectrum of diseases. However, many applications have focused on single tasks and/or single modal data, and thus use networks consisting of a single module. Modules are sections of a complete network, each that carries out a specific task. Multi-module artificial neural networks take inspiration from mammalian brains that process different types of input via dedicated brain regions (e.g. optical information by the visual cortex, sound, and language via Broca's and Wernicke's areas) and subsequently integrates them into a unified representation of our surroundings. While much work has been done on networks with a single module, significantly less work has focused on multi-module networks. Such networks are especially important in clinical problems that require the integration of multi-modal information and/or extracting multiple representations from the same input to provide high predictive performance. This dissertation corrects biases in multi-module networks for clinically relevant neurological problems. It develops novel hyperparameter search strategies with significantly improved performance of multi-module networks for diagnoses and prognoses, and adds uncertainty quantification to empower multi-module mixed effects deep learning models with the ability to produce statistically meaningful measures of covariate significance and principled probabilistic prediction confidence.
Concretely, in this dissertation, I demonstrate the ability of multi-module deep learning networks to integrate spatial and temporal information and automate the detection of artifacts in magnetoencephalography (MEG) brain recordings of subjects including control subjects and those with a head injury. Recognizing how important the optimization of the network architecture was to achieve high predictive performance motivated the development of a novel module adaptive hyperparameter optimization (MA) hyperparameter search framework, which increases the efficiency of architecture optimization of multi-module networks. This approach is demonstrated to identify more optimal architectures when compared to other search strategies and to significantly increase the predictive performance of Alzheimer's and Parkinson's disease prognoses. Finally, I empower mixed effects deep learning (MEDL) models, which explicitly use multi-module networks, with uncertainty quantification allowing for the calculation of fundamental statistical metrics of model fit, covariate coefficient estimation, and prediction confidence. This model is then applied to predict which subjects will convert from mild cognitive impairment to Alzheimer's disease
Regulation of Pumilio RNA Binding Proteins by Long Noncoding RNA NORAD
Get PDFThe mammalian genome is extensively transcribed and encodes thousands of long noncoding RNAs (lncRNAs). Defining the mechanism of action of lncRNAs has been a critical challenge and priority for understanding their biological functions. An important example of this is the lncRNA NORAD, which is a highly conserved lncRNA that is required for maintaining genome stability in mammals. Work in this thesis begins by examining the physiologic function and molecular mechanism of NORAD in human cells. We clearly demonstrate that NORAD localizes predominantly to the cytoplasm where it functions by binding to and inhibiting PUMILIO (PUM1 and PUM2) RNA binding proteins (RBPs). Thorough dissection of the functional domains of NORAD establish the essentiality of PUM binding to NORAD for its function in maintaining genome stability. We further examine the mechanism by which NORAD is able to efficiently sequester and regulate PUM proteins. Through a multidisciplinary approach involving biochemical, microscopy, and mutational analysis experiments, we uncover how NORAD sequesters a super-stoichiometric amount of PUM proteins into novel liquid-like membraneless organelles through a multivalency-induced phase separation mechanism. Moreover, we provide a molecular understanding of the properties of NORAD that promote PUM phase separation, revealing new principles of lncRNA function and demonstrating the physiologic importance of RNA-driven phase separation as a regulatory mechanism in mammalian biology. Our findings, together with the widespread repetitive architecture of lncRNAs, suggest that the phase separation principles we establish may be broadly utilized for lncRNA-mediated regulation
Characterization of [beta]-glucuronidase for Enzyme Replacement Therapy in DYT6 Dystonia
Get PDFThe 61st Annual Medical Student Research Forum at UT Southwestern Medical Center (Tuesday, January 31, 2023, 3-6 p.m., D1.600)Each year the Medical Student Research Program awards students for the best oral presentation and the best poster presentation as judged by faculty across campus. This author received an award as one of the best poster presentations at this forum.BACKGROUND: Dystonia is a debilitating disorder defined by sustained involuntary twisting movements. The current symptomatic treatments for dystonia offer only modest efficacy but numerous side effects. Dominantly inherited, loss of function mutations in the THAP1 transcription factor cause DYT6 dystonia (DYT-THAP1). THAP1 modulates the development of oligodendrocyte progenitor cells (OPC) by regulating the catabolism of glycosaminoglycans (GAGs), a crucial component of the extracellular matrix. The loss of THAP1 within OPCs directly reduces GAG-catabolic lysosomal enzyme β-glucuronidase (GusB) causing the accumulation of GAGs that inhibit their own maturation to myelinating cells. The result is severe dysmyelination during early CNS maturation and impaired neurodevelopment. Genetic overexpression of GusB rescues the maturation deficits and CNS myelination in THAP1 deficient mice raising the critical question of whether β-glucuronidase enzyme replacement could restore myelination in THAP1 null mice.
OBJECTIVE: Characterization of β-glucuronidase for enzyme replacement therapy (ERT) in DYT6 dystonia models.
METHOD: To establish a symptomatic model of DYT6 dystonia, we utilized a Cre/LoxP based Thap1 conditional knockout mouse model ("THAP1-NCKO") and a GusB transgenic mouse line ("GusB-TG"). We evaluated the enzymatic activity and biodistribution of GusB in the CNS using biochemical and histochemical assays.
RESULTS: We determined that THAP1-NCKO mice had lower GusB activity than their control counterparts. Interestingly, the activity of GusB is higher in adult (P60) mice compared to juvenile (P30) mice. Visualization of GusB activity showed that distribution of GusB was highest in white matter tracts. We showed that mice with THAP1- related deficits experienced a significant reduction in GusB activity within white matter tracts but not in other surveyed GusB positive brain areas.
CONCLUSIONS: Age differentially affects CNS GusB enzymatic activity in a murine model of DYT6 dystonia. GusB enzyme exhibits a distinct biodistribution that varies regionally. White matter tracts experience more severe defects with THAP1 loss. Our results provide insights into the specific locations where GusB activity is deficient and highlight the importance of a "critical period" in which genetic insults have long lasting neurodevelopmental implications.Southwestern Medical Foundatio
Building a Methodological Framework for Cell Fate Engineering
No full textCell fate engineering has become an area of intense research in the last fifteen years. A useful framework of cell fate engineering should include three pillars: the discovery of new cell fate-reprogramming cocktails of factors, the evaluation of engineered cells, and the revelation of underlying molecular mechanisms. One major challenge has been the lack of a scalable screening approach in vitro for the performance of reprogramming cocktails. This limits the speed of discovering new cocktails that can efficiently reprogram diverse cell types. Such new cocktails are needed to unleash the full applicational potential of engineered cells in regenerative medicine, disease modeling, and drug discovery. Another challenge is that despite the advantages of in vivo reprogramming, such as more efficient and mature fate conversion, the underlying gene programs, and thereby the molecular mechanisms, have been largely unknown. This is in large part due to the difficulty of specifically isolating and analyzing reprogrammed cells, without contamination from their endogenous counterparts. To address these, in this thesis, I first develop Reprogram-seq, a method that screens thousands of transcription factor cocktails for their reprogramming performance by single-cell perturbation screens. Reprogram-seq found a cocktail of three factors that efficiently and functionally reprograms fibroblasts to epicardial-like cells. Thus, Reprogram-seq accelerates rational cell fate engineering. Next, I performed single-cell transcriptomic analysis of in vivo neurogenesis induced in astrocytes by a novel reprogramming factor, DLX2. This is enabled by a lineage tracer that highly specifically tracks all cells reprogrammed from astrocytes. My analysis reveals that DLX2 induces a neural stem cell-like behavior, transitioning from quiescence to activation, proliferation, and neurogenesis. Gene regulatory network analysis and mouse genetics identify and confirm key nodes mediating DLX2-dependent fate reprogramming. Therefore, this study dissects the gene programs of in vivo reprogramming with single-cell transcriptomics and paves the way for applying Reprogram-seq in vivo. Together, my thesis research has demonstrated that single-cell omic technologies accelerate the discovery of new reprogramming cocktails, streamline the transcriptional evaluation of engineered cells, and dissect gene programs that underlie reprogramming, contributing to all three pillars of the framework. I expect these methodologies to be generalizable to and useful for other cell fate engineering scenarios
43: Ode to the Original Hoonigan
Get PDFThe author submitted this entry in the Open Verse Poetry category (Professional division) for the 2023 On My Own Time (OMOT) Art Show.As a long-time action sports fan, I was deeply saddened by the death of Ken Block earlier this year. His influence transcended snowboarding and rally car racing, and I wanted to pay homage to him. This was published in The New Verse New
Investigating the Impacts of Area Deprivation Index in Rheumatoid Arthritis
Get PDFThe 61st Annual Medical Student Research Forum at UT Southwestern Medical Center (Tuesday, January 31, 2023, 3-6 p.m., D1.600)BACKGROUND/PURPOSE: Rheumatoid arthritis (RA) is associated with increased morbidity and mortality, particularly if RA is poorly controlled. The effects of socioeconomic deprivation have not been well studied in the context of RA. We analyzed the impact of the Area Deprivation Index (ADI) on disease activity in RA.
METHODS: We conducted a retrospective analysis of RA patients, defined by ICD-10 codes, seen at the UTSW and Parkland clinics in the past 5 years and are 18-89 years old. We collected age, ethnicity, race, gender, insurance plan, primary language, address, RA medications, MyChart engagement, primary care physician presence, ED/Inpatient visits over 5 years, RA disease activity and functional scores (RAPID3, HAQ), Charlson comorbidity index (CCI), and CVD presence. ADI was used as a proxy for socioeconomic deprivation and was assigned using 9-digit zip codes. Patients were divided by the upper ADI quartile vs lower ADI quartiles and matched by gender, race/ethnicity, age and CCI for propensity score analysis. Two-sample t test and Chi-square test were conducted for final group comparisons.
RESULTS: 862 patients from UTSW and 1320 patients from Parkland were assessed in the final analysis. The median national ADI score was 45 and 72 for UTSW and Parkland respectively (range 1-100). In the UTSW cohort, those with high ADI scores (more deprived) had significantly higher RA disease activity (RAPID3: 11.9±7.1 vs 13.8±7.1, p<0.0001) and RA functional impairment (HAQ: 0.77±0.7 vs 0.94±0.7 p<0.001), were more likely to have cardiovascular disease (p<0.05), had lower MyChart utilization (p<.0001), more emergency department visits (p<0.05) and had lower jakinib use (p<.0001) compared to those with low ADI scores (less deprived). There were also statistically significant differences in insurance type (p<.05) and smoking status (p<.01). In the Parkland cohort, there were no significant differences between high ADI and low ADI groups in any of the measured variables.
CONCLUSION: We found significant differences in RA disease activity and function in patients from more socioeconomically deprived areas only in the UTSW system. The absence of these differences in Parkland patients raises important questions as to whether certain hospital specific factors influence the role ADI plays in various health outcomes. Identifying the discrepancies between the two hospital systems may elucidate areas of improvement for patient care.Southwestern Medical Foundatio
On Cholesterol Transport Between Membranes
Get PDFThe file named "TRINH-PRIMARY-2022-1.pdf" is the primary dissertation file. One (1) supplemental PDF file is also available and may be viewed individually.The studies described in this dissertation focus on investigation of the pathways for transport of cholesterol from one organelle to another in animal cells. Cells have evolved elaborate transport mechanisms to assure an optimum cholesterol content within their membranes. Dysregulation of cholesterol transport causes common diseases, including atherosclerosis. The major source of cellular cholesterol comes from Low Density Lipoprotein (LDL). When plasma membranes are low in cholesterol, cells produce LDL receptors which bind LDL and mediate its uptake by endocytosis and its delivery to lysosomes. Within lysosomes the cholesteryl esters of LDL are hydrolyzed. The free cholesterol binds to a soluble lysosomal protein called Niemann Pick C2 (NPC2) which delivers it to a membrane-embedded protein called NPC1 which inserts the cholesterol into the lysosome membrane. From there the cholesterol moves to the plasma membrane (PM) through a pathway that is unknown. When the PM becomes saturated with cholesterol, any excess is transported to the endoplasmic reticulum (ER) to repress production of LDL receptors and to be stored in lipid droplets. The work described here 1) showed that triazole antifungal drugs inhibit lysosomal cholesterol export by binding to the membrane domain of NPC1 2) used itraconazole to solve the crystal structure of NPC1 at 3.3Å, 3) revealed that cholesterol is transported out of lysosomes through interactions between two or more NPC1 molecules, and 4) utilized CRISPR-Cas9 whole-genome knockout screens to identify all the genes involved in the transport and uptake of LDL cholesterol. From these screens in the latter study, we discovered that a specific phospholipid, phosphatidylserine (PS), is required for PM-to-ER cholesterol transport. These studies provide supporting evidence towards a vision of one-way directional transport of LDL-derived cholesterol from lysosomes to the PM to the ER
A Boy Who Lived Down the Street
Get PDFThe author submitted this entry in the Open Verse Poetry category (Amateur division) for the 2023 On My Own Time (OMOT) Art Show.Reading about incidents of police brutality against young Black men, I began to remember a childhood neighbor, the son of Nigerian immigrants next door.
We played in our backyards every day one year, then his family moved away. I've since forgotten his name, and it dawned on me that every time I see a news item about a young Black man being hurt - it could be that boy I once played with
Sibling Relationships in Adolescent Eating Disorder Treatment
Get PDFThis study explored the quality of sibling relationships amongst a sample of adolescent patients receiving treatment for an eating disorder. Despite the importance of siblings in adolescents' lives and the utilization of family-based treatment (FBT) as the gold standard treatment, sibling relationships are often overlooked. When sibling relationships have been investigated in the literature, the story has been one-sided, with patient perspectives being omitted. We sought to understand, through the eyes of the patient, the overall quality of the sibling relationship on four factors: warmth/closeness, relative status/power, conflict, and rivalry, and ways in which the onset of the eating disorder changed the relationship regarding these four factors. Of particular interest was discovering specific behaviors, interactions, and aspects of the sibling relationship that patients felt contributed, either negatively or positively, to their eating disorder and recovery efforts. We also hoped to provide patients the opportunity to share ideas or suggestions that could better support their recovery efforts. Methods: Semi-structured individual interviews (N = 14) were completed with adolescent patients. Data were analyzed using a hybrid inductive and deductive approach and thematic content analysis. Results: After the onset of the eating disorder, participants expressed decreased warmth and closeness, increased conflict, and increased competition for family resources. Siblings were often one of the first in patients' lives to suspect the eating disorder. Helpful aspects included greater understanding of their siblings' condition and increased ability to relate to their siblings' struggles, sibling motivation, and involvement in treatment. Comparison, competition, and invalidating or triggering comments were unhelpful. Patients suggested implementation of sibling education and increased involvement in treatment. A common thread throughout this study is the patients' desire to be fully seen, heard, and understood, which they identify to be a critical aspect of achieving recovery. Conclusions: Findings illustrate the importance of sibling relationships among adolescents receiving treatment for an eating disorder. We demonstrate that siblings can be important resources in prevention and early identification efforts. We echo the voices of our patients and advocate for increased sibling education and sibling involvement in treatment, in hopes that our efforts can better support these patients and their families in their journey to recovery