Environmental and Occupational Health Sciences Institute
Rutgers University Community RepositoryNot a member yet
58345 research outputs found
Sort by
Towards controllable generative models
In recent years, generative artificial intelligence (GenAI) has experienced transformative advancements. These ground-breaking generative foundation models have demonstrated remarkable capabilities yet present new challenges. This dissertation delves into the dual aspects of controllability and efficiency in generative models, addressing their potential and limitations. Despite significant progress, generative models often struggle with intricate details and complex structures, such as accurately rendering human hands or adhering to physical laws in video content. These challenges underscore the need for models that can handle complex structures, obey physical principles, and simulate realistic interactions. The research focuses on enhancing controllability, defined as the ability to direct and constrain model outputs to meet specific requirements or user expectations. This is particularly critical in precision-demanding fields like scientific research, engineering design, and healthcare. The dissertation employs a multifaceted approach, including conditioning, model editing, and diffusion-based optimization, to improve the controllability and efficiency of generative models. Conditioning mechanisms explored include categorical labels, preference data, and multimodal inputs, with a shift towards fine-tuning foundational generative models for more complex signals. Model editing and personalization are addressed through Parameter-Efficient Fine-Tuning (PEFT) techniques including SVDiff and SODA.Furthermore, the dissertation investigates diffusion-based optimization for tasks like image editing and 3D human mesh recovery.Ph.D.Includes bibliographical reference
Molecular simulations based pore structure characterization of nanoporous materials
Structural characterization of nanoporous materials in terms of porosity, surface area, pore volume, and pore size distribution is traditionally performed using gas adsorption measurements. This dissertation aims at developing new methods of interpretation of experimentally measured adsorption isotherms using molecular dynamics (MD), Monte Carlo (MC) simulations, and density functional theory (DFT). We focus on theoretical studies of adsorption in two types of nanoporous materials: metal-organic frameworks (MOFs) and organic fractions of shales. First, a new method of pore structure characterization of MOFs is developed based on comparison of experimental isotherm to a kernel of in-silico generated fingerprint isotherms. This method provides valuable information regarding the quality of practical samples in terms of degree of crystallinity and accessibility of pore compartments. An algorithm is devised for calculating the fingerprint isotherms on any MOF structure. Second, mesocanonical ensemble Monte Carlo (MCEMC) is implemented in two open-source software, LAMMPS and RASPA to study adsorption phase transformation like pore filling and capillary condensation. MCEMC method is applied to predict the adsorption isotherm in a series of practical MOF materials. Third, to study the effects of solid framework flexibility a hybrid MD/MC simulation technique has been developed and implemented. Fourth, a multiscale approach is developed by combining MC, DFT, and a macroscopic thermodynamic theory to predict adsorption within kerogen microporous matrix and embedded mesopores accounting for the structure flexibility. The methods developed enhance the understanding of the pore structure and adsorption properties of nanoporous materials and have potential practical applications for pore structure characterization, prediction adsorption properties and ultimately computer-guided design of novel nanoporous materials with advanced properties.Ph.D.Includes bibliographical reference
Regioselective cycloadditions on fullerenes and endohedral metallofullerenes
Over the past few decades, fullerenes and endohedral metallofullerenes (EMFs) have attracted increasing attention for their potential applications of nanomaterials and biomedicine. Chemical functionalization reactions provide fullerene derivatives with substantially improved solubility and modified chemical and physical properties, broadening the scope of applications. Due to the formation of relatively stable cycloadducts in good yields and easy access of reactants, cycloadditions have been one of the most intensively used reactions in the synthesis of fullerene derivatives. This dissertation explores the chemical reactivity and regioselectivity of fullerenes and EMFs in cycloaddition reactions.Chapter 1 provides a general introduction to the fundamental concepts of fullerenes and EMFs, outlining their discovery and structural characteristics. Next, it describes the most common cycloaddition reactions used for the functionalization of fullerenes, including [2+1], [3+2], and [4+2] cycloadditions. Furthermore, it introduces the production, classification, and functionalization of EMFs, with a focus on their chemical reactivity, regioselectivity, and also the challenges associated with multi-functionalization. The chapter concludes with a summary of the current limitations hindering the broad use of fullerenes and EMFs, emphasizing the need for further research to overcome these challenges and unlock the full potential of these materials.
In Chapter 2 and 3, an inverse-electron-demand Diels-Alder reaction was developed on both fullerenes (C60 and C70) and EMFs (Sc3N@C80 and Lu3N@C80). The oxidative [4 + 2] reaction of o-phenylenediamine-derived disulfonamides with C60 or C70 was investigated first in Chapter 2, which we found to be an oxidation reaction followed by an IEDDA reaction. Electron withdrawing groups (EWGs) in the disulfonamides were shown to facilitate the reaction experimentally and computationally, which is the opposite of most fullerene cycloadditions. Indeed, a [4 + 2] adduct on the [5,6]-junction of C70 was isolated as a major product for the first time, which was confirmed by NMR and X-ray diffraction. DFT studies revealed that the dd-[5,6]-C70 adduct is a kinetically favored product.
Next, in Chapter 3, the scope of this reaction was extended to M3N@C80 (M=Sc or Lu). In this reaction, EMFs exhibit higher reactivity compared to empty fullerenes due to the inverse electron demand. Moreover, a complete retro-reaction can be achieved upon thermal treatment. The distinctive reactivity and reversibility of the IEDDA reaction enable a new concept in the chemical separation methodology for EMFs, where the desired EMFs that typically “remain last” in traditional chemical separation methods “react first” and are regenerated by retro-additions. The new chemistry and processing approach provide a key step towards the various applications of fullerenes and EMFs.
In Chapter 4, a new multi-component reaction of alkyl isocyanide, dimethyl acetylenedicarboxylate (DMAD) or terminal alkynes with C60 or Lu3N@C80 were investigated. The three-component cycloaddition reaction of C60, alkyl isocyanide, and DMAD is firstly investigated to afford cyclopent-2-en-1-imino- and ketenimine methano-[60]fullerene derivatives, as confirmed by X-ray single-crystal analysis. Hydration of the ketenimine methanofullerene in the presence of acid gives the corresponding fullerene amides. Additional versions of alkynes have been tested as reactants under similar conditions, which suggests that the reaction requires an ester group on one side and a substituent providing both inductive and resonance effects on the other side of the triple bond. The reaction of t-butyl isocyanide (tBuNC), EWG-bearing terminal alkynes or DMAD with Lu3N@C80 was then investigated. A series of unexpected [6,6]-open metallofulleroids were obtained and fully characterized. The distinct reactivity and selectivity difference between EMFs and empty fullerenes was explored. The difference underscores the influence of the encapsulated metal cluster on the reaction pathway, leading to products with unique electronic and structural characteristics.
In Chapter 5, a highly regioselective bis-addition was achieved by the isocyanide-induced reactions using a two-step strategy. The reaction of tBuNC, terminal alkyne and Lu3N@C80 resulted in a [6,6]-open monoadduct in the first step. The reaction of tBuNC and DMAD was performed on the resulted monoadduct. A series of bis-adducts was obtained as the major products and characterized by HPLC, MS, NMR and X-ray crystallography. Surprisingly, the second addition occurs on the adjacent site of the previous addition site, resulting in a [6,6]-closed structure. This addition shows unique regioselectivity being the opposite of previously reported cases. The computational study on the plausible mechanism and driving force behind the bis-addition is on-going.
Lastly, in Chapter 6, a novel synthetic platform "metallobuckytrio" (MBT) was introduced, which enabled the precise and modular development of water-soluble EMF derivatives. The synthesis of the MBT framework was detailed, which combines a C60 hexakis-adduct with two EMF monoadducts, allowing for controlled addition of functional groups and enhanced solubility. By incorporating polyethylene glycol ligands, the derivatives maintain the electronic and photophysical properties of the EMF cage, providing a promising basis for future development of bio-specific rare-earth element drugs.Ph.D.Includes bibliographical reference
The development and validation of an assessment of romantic relationship boundaries in young adults
Relationship boundaries are a key mechanism through which young adults navigate their intimate relationships, setting terms for how to treat a partner, how they expect to be treated, and the extent to which they allow a partner’s thoughts, feelings, and actions to influence them (Altman, 1975; Cook, 2015). However, there are no validated measures that directly assess relationship boundaries, making it challenging to systematically study their impact on relational outcomes. To fill this gap in the literature, I developed novel assessments of romantic relationship boundaries, the Romantic Relationship Boundaries Scales (RRBS), and validated these scales in a population of young adults ages 18-26. Specifically, I employed a novel integration of attachment theory, complex trauma theory, and family systems work, conceptualizing romantic relationship boundaries in young adulthood as shaped by child experiences of relational safety or trauma in primary caregiver relationships. Drawing from this conceptualization, I used a mixed methods approach to develop and refine an initial item pool of questions that assess boundary attitudes, behaviors, and styles. Subsequently, I used exploratory and confirmatory factor analyses and tests of reliability and validity to revise this item pool and validate four unidimensional scales, each reflecting a different boundary style: Clear, Porous, Protective, and Controlling. Next, in a series of exploratory analyses, I examined if scale scores significantly differed by race, ethnicity, gender identity, sexual orientation, and length of relationship, and subsequently tested the relationship between scale scores and theoretically relevant covariates (i.e., child abuse, relationship quality, and dating violence (DV)) using ordinary least squares regression and logistic regression. Across analyses, RRBA scales demonstrated strong content validity, convergent validity, discriminant validity, and internal consistency (α ≥ .83). Moreover, boundary styles were significantly correlated with theoretically relevant outcomes as predicted, such that: (1) Clear boundaries were associated with higher relationship satisfaction (p<.001), decreased odds of DV victimization (p<.05) and DV perpetration (p<.01), and not having experienced abuse as a child (p<.01); (2) Porous boundaries were associated with lower relationship satisfaction (p<.001) and increased odds of DV victimization (p<.001); (3) Protective boundaries were associated with lower relationship satisfaction (p<.001), increased odds of DV victimization (p<.001), and having experienced abuse as a child (p<.001); and (4) Controlling boundaries were associated with lower relationship satisfaction (p<.001), and increased odds of DV victimization (p<.001) and DV perpetration (p<.001). Creating and validating an assessment of romantic relationship boundaries opens new avenues for empirically studying the role that relationship boundaries play in a range of psychosocial outcomes for young adults.Ph.D.Includes bibliographical reference
Neckpinch singularities of Ricci flow
We use asymptotic analysis techniques to analyze Ricci flow asymptotic to a cylinder and prove two results. First, we derive the precise asymptotic behavior of a compact rotationally symmetric Ricci flow near a singularity modeled on the round cylinder . Second, we derive partial results about the shape of a 4-dimensional steady gradient Ricci soliton with tangent flow equal to .Ph.D.Includes bibliographical reference
Control of behavior by cocaine-sounds encoded in the auditory brainstem and a potential role for the locus coeruleus
Cocaine addiction is a chronic relapsing disorder characterized by the inability to regulate cocaine intake or discontinue use, despite negative health or social consequences. Over the course of initial cocaine use and the transition to addiction, associative-learning processes link sensory stimuli and drug taking to underlie the formation and expression of sensory-cued cocaine memories. Such “maladaptive” memories can initiate, control, and maintain maladaptive stimulus-initiated addiction behaviors like drug seeking and taking, or by inhibiting or otherwise disrupting naturally-motivated behavior (e.g., food and water consumption). Animal models of cocaine conditioning have identified cocaine-dependent plasticity and stimulus encoding in the mesocorticolimbic system as central to the dysregulation of motivation observed in cocaine addiction. However, the neural basis of stimulus control of behavior by cocaine-associated sensory cues remains largely unexplored, despite an extant body of literature on experience-dependent plasticity in primary sensory systems, e.g., like the auditory system, which can encode learned stimulus (sound cue) associations.This thesis reports a novel auditory cocaine conditioning (AuCC) paradigm in rats that blends standard auditory fear conditioning (AFC) and cocaine conditioned place preference (CPP) paradigms, in which cocaine is paired with exposure to a sound stimulus, producing cocaine-sound associative memories. AuCC was used to investigate the stimulus control properties of the cocaine-paired sound on multiple behavioral and neurobiological assays of auditory and neuromodulatory system function in the brainstem. The specificity of effects to cocaine were determined between-subjects conditioned with saline instead of cocaine, or within-subjects in rats conditioned with multiple tones, one paired with cocaine and another paired with saline. The specificity of effects to the stimulus features of the cocaine-paired sound were determined by comparing responses to paired, unpaired, or novel sound stimuli. Several key findings are reported that link cocaine-sound controlled behavior to neuroplasticity in the auditory brainstem and noradrenergic locus coeruleus (LC) activity that together may dynamically encode stimulus features of cocaine-associated sounds.
First, AuCC successfully produces cocaine-sound associations, revealed in multiple forms of cocaine-sound control over spontaneous and motivated behaviors. Second, AuCC induces multiple forms of plasticity in auditory brainstem physiology that include general increases in evoked thresholds in the auditory brainstem response (ABR) across sounds, as well as stimulus-specific changes to sound-evoked amplitude and latency unique to cocaine-paired sounds. Third, cocaine-paired sound cues uniquely activate the LC while animals are in a drug-free state and this activation correlates with the cue-evoked modulation of spontaneous behaviors. Altogether, the significance of the findings herein indicates a role for the auditory brainstem and LC in associative learning processes underlying the control of natural behaviors by cocaine-associated sounds.
These findings are consistent with an overarching model wherein sensory-drug associations are encoded by neuroplasticity in sensory systems and mediated by neuromodulator systems that enable cues to gain control of behavior and subserve addiction phenotypes.Ph.D.Includes bibliographical reference
Developing novel strategies to regulate metabolic resource distribution for enhancing microbial bioproduction
Microbial biosynthesis has been well recognized as a sustainable and environmentally friendly approach for production of various products and has attracted significant attention due to its applicability in different industries. Advances in metabolic engineering and synthetic biology have accelerated the development of novel microbial biosynthesis methods to convert renewable feedstocks into high-value products. However, one of fundamental challenges in microbial biosynthesis, the competition for resources between cell growth and bioproduction, is still largely unresolved. This dissertation focuses on developing innovative strategies to balance the allocation of cellular metabolic resources between cell growth and bioproduction to effectively enhance the biosynthesis of desired products. First, a novel strategy of antimicrobial peptide (AMP)-mediated global metabolism regulation was developed by expressing an AMP in E. coli to inhibit protein synthesis process and thus repress global metabolism. Compared with the traditional method of over-expression of the biosynthetic pathway genes or the method of direct addition of an exogenous antibiotic, this strategy demonstrated significant advantages in enhancing bioproduction of target products. Further improvements were achieved by optimizing the conditions for using AMP. The successful application of AMP-mediated global metabolism regulation across different bioproduction systems highlighted its effectiveness and broad applicability in microbial biosynthesis. Systematic characterization of the engineered E. coli, including the cell growth, bioproduction, protein synthesis, and energy, indicated that this strategy effectively reallocated metabolic resources among key destinations to enhance the desired bioproduction. Metabolomics analysis further revealed that the use of AMP delayed carbon source uptake and significantly increased metabolite pool size, which benefitted the bioproduction with higher pathway enzyme expression than the background metabolism. This strategy was further integrated with co-culture engineering technology to enable the over-production of a pathway intermediate for an E. coli-E. coli co-culture as well as more rational allocation of resources between the co-culture members, which resulted in significant improvement of the production of the final product. This dissertation also explores another novel strategy to address resource allocation in a co-culture by incorporating a metabolite biosensor into the co-culture to dynamically regulate the population composition to meet the need of biosynthesis optimization. Specifically, a biosensor-regulator system responsive to a biosynthetic pathway intermediate was employed in the downstream co-culture strain. Through dynamic regulation of the downstream strain’s growth based on the pathway intermediate accumulation, this system rationally assigned resources to the co-culture strains to balance the pathway intermediate supply and consumption and significantly enhanced the overall bioproduction performance of the target product indigo.
The successful application of the engineering strategies developed in this dissertation, including the AMP-mediated global regulation, the biosensor-regulator system, and their integration with co-culture engineering in various bioproduction systems, provided innovative perspectives to not only address the longstanding challenge in microbial biosynthesis but also facilitate the advances in metabolic engineering and synthetic biology.Ph.D.Includes bibliographical reference
Light-dependent ecophysiological costs and benefits of calcification in the coccolithophore Emiliania huxleyi
Biogenic calcification has been widespread throughout marine organisms since the Cambrian era. In the modern ocean, coccolithophores are the most abundant calcifying algae. Calcification in coccolithophores uses bicarbonate to produce calcite plates known as coccoliths and carbon dioxide is released as a byproduct. Coccoliths facilitate carbon export to depth by ballasting organic matter. In addition, they scatter light to such a degree that blooms of the most widespread coccolithophore species, Emiliania huxleyi, can be viewed from Earth-observing satellites. Despite the considerable amount of attention that calcification has garnered over the years, there are still degrees of uncertainty surrounding the mechanisms and purpose of calcification. There are several hypotheses that attempt to address the ecophysiological role of calcification, from protecting cells from grazers and viral infection, to enhancing photosynthesis by modulating the amount of light entering a cell, and/or by increasing intracellular carbon concentrations. Furthermore, uncertainty remains around the cellular costs and benefits associated with calcification and how they co-scale with light and photosynthesis. The aim of my PhD thesis is to better characterize the ecophysiological costs and benefits of calcification across irradiance.My dissertation is divided into three chapters, each focusing on different aspects of this objective. After a brief introduction:
Chapter 1 demonstrates how calcification enhances photosynthesis. This work uses genetically-related calcified and non-calcified (‘naked’) phenotypes of E. huxleyi grown across different light intensities (25-2000 µmol photons m-2 s-1). I investigated the impact that light intensity has on calcification and the light and dark reactions of photosynthesis using flow cytometric methods, PIC and POC analysis, fluorescence-based biophysical measurements of photosynthesis, quantitative immunoblots, 14C incubations, and stable isotope analysis. I show that coccoliths serve a photoprotective role and that calcification increases the bicarbonate available for organic carbon fixation. I also demonstrate that coccolith shedding rates are modulated by irradiance.
Chapter 2 investigates the molecular mechanisms supporting calcification and assesses what impact calcification has on the expression of metabolic pathways. Through differential gene expression analyses, I identify an alpha carbonic anhydrase that plays a role in calcification. In addition, I find that calcification is associated with downregulation of antioxidant and photorespiration pathways. Expression patterns of ion transporters and a vacuolar H+-ATPase suggest that regulation of ion transport and activity in calcifying E. huxleyi may not occur at the transcript level.
Chapter 3 characterizes how the intracellular partitioning of carbon is impacted by calcification. The results from Chapters 1 and 2 demonstrated that calcification increases the carbon supplied for organic carbon fixation. Using empirical data and a numerical model, I demonstrate that calcification is part of a multipronged carbon concentrating mechanism in E. huxleyi. My findings also suggest that calcification is associated with higher rates of protein synthesis and turnover. Lastly, increased lipid quotas associated with calcification suggest that lipids may play a role in buoyancy maintenance.Ph.D.Includes bibliographical reference
Investigation of the TRPM7 COMPLEX in mouse embryonic stem cells
Transient Receptor Potential Melastatin 7 (TRPM7) is an ion channel and kinase with broad functions in cell physiology, including cellular magnesium homeostasis, cell proliferation, and cytoskeletal remodeling. TRPM7 is permeable to the essential divalent cations Mg2+, Ca2+ and Zn2, which are fundamental to the channel’s function. TRPM7’s kinase has been shown to phosphorylate multiple substrates and when cleaved from the channel, regulate gene expression events in the nucleus. TRPM7 is also the first ion channel shown to be indispensable for embryonic development. Loss of the channel-kinase in mouse and Xenopus laevis, results in embryonic lethality. Despite extensive research on the channel over the past decades, how TRPM7 exerts its many cellular functions, particularly during embryonic development, is poorly understood. Recently TRPM7 was found to assemble in rodent brain with newly identified interacting partners: CNNMs (cyclin and CBS domain divalent metal cation transport mediators), ARL15 (ADP ribosylation factor-like protein 15), and PRLs (Phosphatases of Regenerating Liver). We hypothesize that TRPM7 assembles with these proteins in embryonic stem cells to regulate TRPM7’s diverse functions. To test our hypothesis, we used mouse embryonic stem cells (mESCs) as our model system to investigate TRPM7’s role in early developmental processes. Using quantitative reverse transcription polymerase chain reaction (qRT-PCR) and RNA sequencing (RNAseq), we demonstrate that Cnnm2 and Cnnm3 are the predominant CNNM members in mESCs. Mass spectrometry analysis further reveals that CNNM2 and CNNM3 are the dominant CNNM proteins that interact with the channel in mESCs. TRPM7 knockout (KO) mESCs exhibit impaired proliferation, despite magnesium supplementation in the growth media. Our experiments also demonstrate that TRPM7 KO mESCs display altered gene expression of pluripotency markers, including Nanog and Sox2. Moreover, RNAseq analysis of TRPM7 KO mESCs reveals significant transcriptional changes, with over 500 differentially expressed genes (DEGs). We generated CNNM3 homozygous and heterozygous knockout mESCs and show that they retain viability and express key pluripotency markers. Nevertheless, deletion of CNNM3 from mESC significantly reduces TRPM7 channel activity. Differentiation assays using mESCs demonstrate that TRPM7 KO cells exhibit widespread changes in gene expression and potentially influences lineage specification in mESCs, with TRPM7 KO cells showing upregulation of mesodermal and endodermal markers. Neuroectodermal (NE) differentiation of TRPM7 KO mESCs show altered morphology of TRPM7 KO NE cells and significant changes in gene expression compared to WT NE cells. Mass spectrometry analysis of NE cells identify CNNM4, CNNM3, CNNM2 and ARL15 as the main TRPM7 interacting proteins in NE cells. Together, these findings advance the current understanding of TRPM7’s interactome in early embryonic cells, revealing new insights into TRPM7 and CNNM’s roles in mESC pluripotency, proliferation, and lineage specification.Ph.D.Includes bibliographical reference
Stories lost in time: a case study on the posthumous travels of Iset-Ha
In the Rutgers University Geology Museum (RUGM) collection lies the mummy of a Ptolemaic Priestess named Iset-Ha. A Presbyterian missionary, John G. Lansing brought Iset-Ha to the RUGM after a trip to visit his parents in Egypt. Since her arrival to the United States, Iset-Ha’s life story has largely been a mystery, with very little information on who she was and how exactly she ended up in the United States. This paper aims to take the information that is known of Iset-Ha, including archival materials and piece together an interpretation on what her life and afterlife has been. This paper serves to be a case study in the greater issue of human remains in museum collections.M.A.Includes bibliographical referencesIncludes vit