Environmental and Occupational Health Sciences Institute
Rutgers University Community RepositoryNot a member yet
58345 research outputs found
Sort by
Tissue disposition of cadmium and nutritional metals in mice with a reduced functioning BCRP transporter
No full textThis thesis will explore the tissue distribution of the toxic metal cadmium and its interplay with a prominent efflux transporter. The main transporter covered in this thesis is the BCRP transporter. Otherwise known as the breast cancer resistance protein, the BCRP transporter was shown to reduce the accumulation of cadmium in different types of tissues. Cadmium has no beneficial biological purpose in the body and can result in negative effects such as kidney failure and liver dysfunction. Although the target organs of toxicity are mainly the kidneys and liver, cadmium has also been known to accumulate in the testis, and small intestinal tract. This thesis studies the mouse wild type Bcrp variant and compares it to the Q140K variant, which is genetically modified to mimic the reduced functionality of the human Q141K BCRP transporter in mice, in order to compare how effective they are in eliminating cadmium from tissues. Inductively Coupled Plasma Mass Spectrometry (ICPMS) was used to identify which tissues accumulate cadmium and quantify absolute concentrations in target tissues.
The purpose of this study was to understand the role of the Bcrp transporter in decreasing cadmium bioaccumulation, and to analyze cadmium’s distribution within specific organs in a C57BL/6 male mouse model. In-vitro and in-vivo experiments have helped in understanding that transporters might either reduce or increase cadmium bioaccumulation and toxicity in the body. To test the role of the human Q141K variant in vivo, wild-type C57BL/6J mice were modified by CRISPR to contain the same nonsynonymous polymorphism at position 140 (i.e., Q140K). Tissues from mice with the Q140K Bcrp variant have lower Bcrp protein expression, and therefore reduced transport activity compared to the wild-types. Thirty-two mice were split into control and CdCl2-dosed groups for both WT and Q140K genotypes. The Cd-dosed groups were provided MQ water spiked with 50ppm of CdCl2 for drinking while the control groups were provided regular MQ water for 14 days. The Cd-exposed groups had a significant increase in body weight compared to the control groups. ICPMS was used to quantify cadmium, manganese, zinc, and copper. Results showed a significant increase of cadmium concentrations in the testis, liver, and kidneys of the Q140K mice, while there was a significant decrease in both cadmium in the plasma of Q140K mice. Cadmium showed no significant differences within the small intestinal tract between WT and Q140K mice. Zinc showed a significant increase in the liver of Q140K mice, and a significant decrease in the plasma of WT mice. Manganese showed a significant decrease in the liver and duodenum for Q140K, and a significant decrease in the liver for WT mice. Selenium showed a significant in both plasma and testis, as well as a significant increase in the kidneys for Q140K. There were no significant changes for copper. Multiple transporters are involved in the uptake and efflux of cadmium. RNA sequencing was used to compare the Bcrp transporter to other proteins previously studied that might contribute to cadmium clearance or cadmium toxicity. RNAseq data from liver samples were analyzed and showed an increase of cadmium binding proteins metallothionein, lipocalin-2, and transferrin in the cadmium-treated Q140K mice compared to the cadmium-treated WT group. There was a decrease in the mRNA levels of the heat shock proteins Hspa1a, Hspa1b, and Hspa4l in the cadmium-treated Q140K group compared to the WT. Zinc transporters Znt1 and ZIP14 were increased in the livers of Cd-treated Q140K mice, while ZIP10 mRNA decreased compared to the WT group. Quantitative PCR data of metallothionein also show an increase in the cadmium-treated Q140K group and western blot data show a decrease of Bcrp protein and an increase of HO-1 protein in the Q140K group compared to the WT group following administration of CdCl2. Clinical chemistry data showed no significant difference between WT exposed and Q140K exposed groups for BUN, ALT, ALP, creatinine, calcium, phosphorus, albumin, cholesterol, or bilirubin. Glucose has a significant difference with the cadmium exposed WT group having higher levels than the control WT group, and a similar but non-significant trend between the Q140K groups.The BCRP transporter plays a major role in reducing the accumulation of cadmium in the body. The reduced activity in the Q140K variant alters the activity of other transporters that play a role in cadmium transport, and the transport of other metals. With cadmium concentrations higher in the Q140K mice compared to the wild type group, this could mean that humans with a reduced functioning BCRP variant could be at a higher risk of developing cadmium toxicity than other mammals, and more research needs to be done to reduce the negative effects that cadmium accumulation causes.M.S.Includes bibliographical reference
AI-driven assessment and scaffolding for mathematical modeling and explanations during science investigations
No full textDeveloping models, using mathematics, and constructing explanations are three practices essential for science inquiry learning according to education reform efforts, such as the Next Generation Science Standards (NGSS Lead States, 2013). However, students struggle with these intersecting practices, especially when developing and interpreting mathematical models (i.e., graphs and corresponding equations) of scientific phenomena and writing explanations that connect the mathematical representations to their scientific meaning. As such, students require targeted support to address their difficulties while they are working in science inquiry contexts. Thus, recent work with the Inquiry Intelligent Tutoring System (Inq-ITS) virtual lab environment focuses on the design and development of performance-based formative assessments with the goal of automatically assessing and supporting students on fine-grained components underlying mathematical modeling as well as exploring novel techniques for assessing explanation writing. The studies presented in this dissertation were conducted to evaluate the efficacy of the automated scaffolds in the Inq-ITS environment and examine how large language models could be leveraged to assess students’ written explanations about their mathematical models in a way that provides formative data to teachers to guide their instruction. These findings inform how innovative technologies, such as virtual inquiry environments, can be used to assess and support NGSS-aligned mathematical modeling and explanation writing competencies in science inquiry contexts, thus preparing students for future success in mathematics and science.Ph.D.Includes bibliographical reference
Investigating HIV care in the United States: race, gender and policy
No full textBackground: Black Americans are disproportionately affected by HIV. Despite the availability of biomedical interventions that either prevent or manage HIV, disparities remain. Limited research has investigated antiretroviral (ART) adherence outcomes among Black Americans. Further, minimal research understands the role of social determinants of health (SDoH) and public policy in ART adherence and other HIV care outcomes. This three-paper dissertation contributed to HIV empirical literature by examining how race, gender and policy influenced HIV outcomes among U.S. populations. Methods: This dissertation included two empirical quantitative papers investigating the outcome variable, ART adherence, for people living with HIV (PWH). The last paper in this dissertation is a systematic review that investigated the inclusion of cisgender Black women in policy-driven HIV studies. For the quantitative papers, local Medical Monitoring Project surveillance data were examined. Paper 1 included a sample size of 716 adults with HIV living in a Midwestern state with Medicaid expansion. Data were collected between 2018-2021. Logistic regression models were used to estimate racial differences in ART adherence. Additional models estimated the effects of discrimination in healthcare settings and HIV-related stigma on the outcome of interest. Next, paper 2 used difference-in-differences models to estimate the effects of a social determinants of health (SDoH) - driven policy on ART adherence among PWH living in a southern state (n=1,207). Last, the systematic review examined literature considering cisgender Black women’s HIV outcomes in policy research. This paper followed the PRISMA review guidelines and is registered in the PROSPERO database.
Results: In paper 1, results indicated that discrimination in health care settings decreased the ART adherence for PWH, although this was not true for HIV-related stigma. The effects of stigma were explained away when other covariates. Consistently in all models, Black PWH were significantly associated with lower odds of ART adherence compared to PWH in other racial/ethnic groups (p<.05). Results in paper 2 found that PWH in the U.S. South had a 11% higher likelihood (p<.05) of ART adherence in the post-policy years (2019-21), compared to years 2015-18. Analyses stratified by insurance status suggested that factors other than Medicaid policy may be driving this improvement. Further, food insecurity (p<.05) and housing need (p<.05) significantly lowered the likelihood of optimal ART adherence. Last, the systematic review found that few studies in the literature contextualized cisgender Black women’s needs within major U.S. health policies. Among the findings, emergent themes included biomedical intervention uptake and adherence, health insurance, poverty and programmatic barriers to care.
Conclusion: This dissertation documents the impact of discrimination and policy on PWH, particularly for Black PWH. Despite innovative progress in HIV interventions, discrimination is a critical factor in ending the epidemic. Relatedly, although SDoH policies may improve barriers, results indicate that stakeholders should consider the specific needs of Black PWH to ensure meaningful outcomes for the population. The systematic review highlights the needs for further research, among large populations of cisgender Black women, notably in HIV prevention, to understand how federal and policy decision encourage, create, or exacerbate barriers to HIV care.Ph.D.Includes bibliographical reference
Application of the Berry phase to problems in solid-state physics
No full textThe Berry phase, a geometric phase acquired by the wavefunction of a quantum system during cyclic adiabatic evolution, is a fundamental concept in quantum mechanics with profound implications across various fields, including solid-state physics and electronic structure theory. This thesis explores several aspects of electronic structure theory related to the Berry phase, namely surface orbital magnetization, electric polarization switching, and the formation of topological Chern bands in two-dimensional materials. We begin by introducing the foundational concepts and methods used throughout the thesis. Subsequently, we investigate surface orbital magnetization, first identifying macroscopic bound currents running along hinges shared by two surface facets as the corresponding physical observables. We then explore the symmetry conditions for surface magnetization to be well defined and propose a method to compute it using a quantum local marker. Tests on tight-binding models reveal that only one specific marker consistently predicts the hinge currents correctly. Next, we examine electric polarization switching pathways in CuInP2S6, a van der Waals layered ferroelectric material. Using first-principles density functional theory calculations, we explore the rich energy landscape of the system and identify "sequential" paths -- those that break the system's glide mirror symmetry and involve independent Cu ion motion -- as energetically favorable for switching through the monolayers and van der Waals gaps. The combination of such through-layer and cross-gap paths is demonstrated to form a globally insulating cycle resulting in Cu ions shifting into neighboring unit cells, permitting a characterization of CuInP2S6 as a quantized adiabatic charge pump. Finally, we investigate the application of magnetic superlattices, generated by spatially modulated magnetic fields, to bilayer graphene for the purposes of inducing flat topological Chern bands. We explore fields that introduce either no net magnetic flux or a single quantum of flux to the superlattice unit cell, as well as their joint action with electric superlattices. We then propose a method of generating single-flux-quantum magnetic superlattices and concomitant electric superlattices involving charge densities induced on the surface of magnetoelectric materials by the flux vortices of a type II superconductor. We study the effects of these superlattices on the band structure of bilayer graphene.Ph.D.Includes bibliographical reference
Super-resolution and edge-assisted delivery for enhanced video streaming
No full textWith the increased popularity of video streaming platforms, users are spending more time watching and sharing videos. Besides traditional video streaming on flat screens, 360- degree video has also recently gained popularity. The 360-degree video provides immersive experiences and is typically viewed on head-mounted displays, requiring more resources than traditional streams. Because demand for higher-resolution video and more immer- sive experiences scales more quickly than network resources can be provisioned, we now require bandwidth-efficient solutions for video streaming and sharing. In this dissertation, I present our contributions to enhancing video streaming under lim- ited network bandwidth via deep-learning-based super-resolution and edge-assisted stream- ing. Our contributions can be summarized into three major efforts. i) We propose EVASR, a system that uses edge-assisted video delivery to address the limited network bandwidth and the high computation demand of super-resolution. EVASR efficiently uses the shared computation resources available at the edge server by performing super-resolution on frame patches selected by a saliency visual quality optimizer (SVQO), while the remaining patches are upscaled via bicubic interpolation. ii) Unlike traditional 2D videos, visual information in 360-degree videos is naturally represented as pixels on a sphere. We propose a spherical super-resolution (SSR) approach that uses a novel VertexShuffle operation on the mesh. To support viewport-adaptive and bandwidth-efficient transmission of 360-degree video data and save computation, we create a Focused Icosahedral Mesh for the representation of spherical content. iii) To address the uploading bandwidth constraint in 360-degree video live streaming, we propose Salut - an uplink transport system that utilizes novel video projections and separates the static background from the active foreground to minimize uplink requirements. Salut takes advantage of cloud servers for computational offloading tasks, incorporating saliency detection techniques to detect the active foreground objects and stitch the separated streams. Moreover, Salut employs a unique control and communi- cation protocol to enable real-time computational offloading between mobile devices and servers, ensuring a seamless and uninterrupted live streaming experience.Ph.D.Includes bibliographical reference
The gut microbiome and Vitamin B12 status in humans
No full textBackground: Vitamin B12 is an essential nutrient required for hematological and neurological function. B12 deficiency results in megaloblastic anemia and neurodegenerative disease and may contribute to cognitive decline and dementia risk in older adults. Some bacteria among human gut microbiota can synthesize B12 while others require B12 from the diet of the host. Consequently, variation in gut microbiota may influence B12 status in humans. Aims: The purpose of this dissertation is to assess the associations between gut microbiota composition and host B12 status as assessed by serum total B12, holotranscobalamin (HoloTC), methylmalonic acid (MMA), and total homocysteine (Hcy) concentrations.
Methods: A pilot study was first conducted consisting of a secondary analysis of a cohort of 90 adults (44 males/46 females; 20-75y) in Sydney, Australia. A second study was conducted with a cohort of 100 healthy adults (76 females/24 males; 18-81y) recruited in New Brunswick, New Jersey. 16S rRNA gene V3-V4 amplicon sequencing was used to profile gut microbiota at the amplicon sequence variant (ASV) level. Serum B12 was measured by automated chemiluminescence assay, serum MMA by high pressure liquid chromatograhy-tandem mass spectrometry, serum HoloTC by active B12 ELISA kit assay, plasma Hcy by HPLC with post-column fluorescence detection, and plasma short chain fatty acids (propionate, butyrate, and acetate) by gas chromatography-tandem mass spectrometry. In addition, a composite score called combined B12 (cB12) was calculated using a published formula that includes all four blood B12 biomarkers. Microbial guilds were identified based on co-abundance clustering using the Ward clustering algorithm. Random forest regression with feature selection by the Boruta algorithm was used to identify microbial co-abundance guilds that are predictive of measured B12 status biomarker concentrations adjusting for age, sex, body mass index (BMI), and fiber intake in the Australian cohort, and adjusting for age, sex, BMI, race/ethnicity, creatinine, alcohol intake, caffeine intake, fiber intake, and supplemental B12 in the New Jersey cohort. Linear regression was used to assess the associations between the guilds and B12 status biomarkers, and short chain fatty acids (New Jersey cohort only) in univariate analyses and in multivariate analyses adjusting for the covariates listed above for the two cohorts. Mediation analysis was conducted to determine if propionate mediated the associations between guilds and MMA in the New Jersey cohort.
Results: In the Australian cohort, mean ± SD serum B12 and MMA were 389 ± 177 pmol/L and 210 ± 80 nmol/L. Thirty microbial co-abundance guilds were identified. Random forest regression with controlling for age, sex, BMI, and fiber intake, revealed that age and guilds 3 and 15 were predictors of B12 (p=0.027), and guilds 4, 27, and 29 were predictors of MMA (p=0.003). Multivariate linear regression revealed that little of the variance in B12 (adjusted R2=0.038) was associated with the guilds, while a higher amount of variance in MMA (adjusted R2=0.196) was associated with the guilds. In the New Jersey cohort mean ± SD serum B12, MMA, HoloTC, Hcy, cB12 and propionate were 349 ± 143 pmol/L, 161 ± 119 nmol/L, 92 ± 49 pmol/L, 7.75 ± 2.36 µmol/L, 0.65 ± 0.05 and 1.34 ± 0.76 µmol/L, respectively. Twenty-eight microbial co-abundance guilds were identified. Random forest regression with controlling for covariates revealed that supplemental B12 and guilds 14, 15, and 26 were predictors of B12 (p<0.0001); guilds 2, 6, and 14, sex, creatinine and BMI were predictive of Hcy (p<0.00001), and guild 13 was a predictor of MMA and cB12 (p=0.01 and p=0.001). No guilds were predictive of HoloTC. In the multilinear regression, serum B12, HoloTC and cB12 models did not reach significance. The amount of variance in Hcy explained by the model was strong (adjusted R2=0.116) and was strengthened by the addition of covariates (adjusted R2=0.201). Likewise, the amount of variance in MMA explained by the model also was strong (adjusted R2=0.218). Propionate was directly associated with MMA and explained 12.6% of the study sample variance in MMA (R2=0.126; p=0.0003). After controlling for age, sex, B12, creatinine, BMI, and fiber intake, the association with propionate remained significant (p=0.002) with the full model including all covariates explaining 31.1% of the variance in MMA (R2=0.311). No significant associations between butyrate or acetate and MMA were found. The mediation analysis revealed that the mediation effect of propionate of the association between one of the guilds (guild 6) and MMA was 20%, but this mediation effect did not reach statistical significance (p=0.14).
Conclusion: These results indicate that in these cohorts of healthy adults different microbial co-abundance guilds are predictive of total B12, MMA, HoloTC, Hcy and cB12. Moreover, guilds account for a larger percentage of the variance in MMA and Hcy compared with total B12 and HoloTC. Propionate appears to partially mediate the relationship between MMA and at least one guild. This suggests that some of the variance in serum MMA and Hcy in adults is not specific to B12 status. MMA is generally considered to be a specific biomarker of B12 status, but is also influenced by renal function and age, and as we show here variance in the gut microbiome. This variance is mediated by propionate. This has ramifications for interpreting measures of B12 status in healthy adults and may provide insight into the metabolic connections between gut dysbiosis and diseases such as cancer, cardiovascular disease and neurodegenerative disease.Ph.D.Includes bibliographical reference
Application of quantitative modeling techniques to assess drug toxicity and support optimization of drug dosing
No full textAs we continue to seek to improve drug efficacy while minimizing toxicity for both patient benefit and increased economic gain, translational research and the pharmaceutical industry are increasingly relying on more sophisticated quantitative modeling approaches. This paper aims to investigate how these approaches can be utilized, specifically by integrating three key modeling approaches: physiologically based pharmacokinetic (PBPK) modeling, quantitative systems pharmacology (QSP) modeling, and population pharmacokinetic (popPK) modeling. These models offer three unique perspectives to predict the behavior of drugs in biological systems, enabling more efficient and safer development of drugs for clinical use.These models are introduced in Chapter 1, along with an overview of their importance and evolution within the pharmaceutical research community. Chapters 2 to 4 provide examples of model applications: (1) Chapter 2 describes the use of PBPK models for characterizing apixaban disposition and for interspecies scaling. The findings presented in Chapter 2 can help translate animal-based models to humans for predictions of drug distribution in tissues and potential toxicities and adverse events; (2) Chapter 3 uses popPK models to evaluate the dosing of rifampin in pediatric and adult populations and demonstrates how this pharmacokinetic analysis can aid in the optimization of dosing regimens and help characterize pharmacokinetic variability; and (3) Chapter 4 shows how PBPK and QSP modeling strategies can be used synergistically to mitigate cardiotoxicity in cancer treatment. The major outcome of this chapter is the development of a comprehensive modeling strategy for characterizing clinical cardiotoxicity biomarker profiles that can be used to support optimal dosing to reduce cardiotoxicity in breast cancer treatment.
Overall, this paper demonstrates the application of quantitative models in pharmacokinetic research and clinical practice to optimize treatment strategies and provide references for regulatory decisions. The results show that advanced modeling methods can greatly accelerate certain stages of the drug development process, improve therapeutic efficacy, and reduce its adverse reactions, which is a key advance in pharmaceutical science.Ph.D.Includes bibliographical reference
Palliative care, mental health services, and suicide risk among older adults with metastatic cancers
No full textOlder adults with metastatic cancers have high prevalence of psychiatric conditions and symptoms and have increased suicide risk. Health services use such as psychotherapy use and antidepressant medication can reduce suicide risk among individuals identified as at risk. Prior studies have suggested that early palliative care use is associated with reduced suicide risk among Veterans. This dissertation, presented in three papers, aims to understand important factors associated with early palliative care use and psychotherapy initiation, and the role of early palliative care use in suicide risk reduction among older adults with metastatic cancers.Manuscript One uses fee-for-service Medicare data to create a cohort of older adults with newly-diagnosed metastatic cancers, and focuses on trends in early palliative care use and important predictors of early palliative care use among older adults with newly-diagnosed metastatic cancers. Early palliative care use increased in prevalence throughout the study period, and individuals with poorer-prognosis cancers and certain psychiatric diagnoses were likely to receive early palliative care.
Manuscript Two uses the same base cohort to examine rates and predictors of psychotherapy initiation in the year following metastatic cancer diagnosis. Those with new diagnoses of major depressive disorder, anxiety disorders, adjustment disorders, and post-traumatic stress disorder had the highest likelihood of psychotherapy initiation. In addition, men, Black patients, and those older than 80 were more likely than their counterparts to initiate psychotherapy, despite limited evidence that psychotherapy is successful for individuals with advanced cancers who belong to these demographic groups.
Manuscript Three uses the same base cohort as Manuscripts One and Two and examines the effect of early palliative care use on suicide risk for the three years following metastatic cancer diagnosis, over and above the impact of mental health services use. Those who received early palliative care had significantly reduced risk of 3-year suicide mortality, however a propensity score analysis did not find a significant effect. Routinized early palliative care may reduce suicide mortality among older adults with metastatic cancers.
Taken together, these findings highlight the need for increased palliative care workforce and integration of mental health specialists in palliative care teams to aid in management of psychiatric symptoms among older adults with metastatic cancers.Ph.D.Includes bibliographical reference
Cryogen-free low temperature high magnetic field ultra-high vacuum modular scanning tunneling microscope for the exploration of van der waals materials and nano-devices
No full textScanning probe microscopy (SPM), particularly scanning tunneling microscopy (STM), has become a critical tool for locally probing material properties at the atomic level. STM continues to evolve to meet diverse and challenging research demands. Advancements have enabled high-resolution imaging and detailed analysis of complex materials and quantum states at low temperatures (LTs). However, challenges arise with LT-STM operation due to helium shortages, affecting the availability and cost of liquid helium which is essential for maintaining LTs. To address these challenges, cryogen-free closed-cycle cryostats (CCCs) have emerged, offering a sustainable alternative to liquid helium systems, since helium gas is recycled and liquid helium is not consumed. However, CCCs introduce high levels of mechanical vibration and acoustic noise, complicating integration with STM systems. Additionally, the need for ultra-high vacuum (UHV) environments and high magnetic fields further complicates STM design, particularly for experiments involving μm-sized 2D material sample devices. This thesis presents the design, development, and performance of an innovative STM head that integrates modularity, flexibility, stability, and compact size. The STM head is equipped with a quick-connect socket that is matched to a universal connector plug, enabling it to be transferred between systems. This head can be introduced into a vacuum system from room temperature and seamlessly transferred to various sites equipped with the connector plug, permitting multi-site STM operation is a variety of experimental conditions, including broad temperature ranges, ultra-high vacuum, high magnetic fields, and closed-cycle pulse-tube cooling. The STM’s compact size is achieved by a novel nested piezoelectric coarse walker design which allows for large orthogonal travel in the X, Y, and Z directions, ideal for studying both bulk and thin film samples ranging in size from mm to μm. Its stability and noise tolerance are demonstrated by achieving atomic resolution in ambient conditions on a laboratory desktop with no vibrational or acoustic isolation.
This thesis introduces the design, construction, and performance of a cryogen free, UHV, LT, and high magnetic field platform for modular STM operation at variable temperatures, a base temperature of 4K, and magnetic fields up to 9T. This platform incorporates a novel internal vibration isolator to mitigate vibrations and a triple wall vacuum design to achieve low base temperatures. The CCC is incorporated into a UHV platform that supports in-situ sample and tip conditioning and alignment. Multi-site operation without breaking vacuum is facilitated by the quick-connect STM head and vacuum transfer methods, notably a low-profile vertical transfer mechanism.
This thesis presents an internally mounted flexible mechanism that replaces large externally mounted vertical translation stages by utilizing collapsible components and gravity assisted tools to insert and extract the STM head into the LT region. This design substantially reduces the size of the vertical transfer mechanism which enables installation in standard laboratory spaces. The design eliminates also excess pumping volume and reduces susceptibility to vibrational noise.
The combined innovations presented in this thesis significantly enhance STM capabilities, providing a platform for studying a wide range of samples under various experimental conditions. To demonstrate the system’s capabilities, and exploration of van der Waals heterostructures was conducted. This thesis presents the study of magic angle twisted-bilayer graphene (TBG) nearly aligned to hexagonal boron nitride (hBN). The stacking dependence and rotational misalignment of the layers is explored through STM topography and spectroscopy of the moiré superlattice to explore the formation of a moiré intercrystal.Ph.D.Includes bibliographical reference
Machine learning-based prediction on heat transfer enhancement in complex thermal systems
No full textThe complexity of interactions between multiple factors and heat transfer enhancement creates a strong desire to investigate new approaches as an alternative to traditional numerical and experimental methods. This study aims to evaluate the applicability of machine learning techniques in identifying the complicated nonlinear patterns in huge dataset and predicting the performance of heat transfer enhancement systems, including microscale pin fin heat sinks and nanofluids. Various types of machines learning algorithms, including multilayer perceptron, random forest, light gradient boosting machine, extreme gradient boosting and stacking methods are employed to develop a universal predictor applicable under various operating conditions. Experimental data acquired from the literature were employed to train the prediction model. The performance of the trained machine learning models was tested using the mean squared error and the coefficient of determination. The influence of each input variable on model development was examined to identify key features. Information gain is introduced and calculated for determining the importance of parameters in prediction. External validation is performed with an additional unseen dataset to further assess the applicability of the selected models across different experimental data points. The results demonstrate that machine learning methods can provide broad applicability and excellent predictions of heat transfer enhancement of microscale pin fin heat sinks and complex nanofluids with minimal prediction error.Ph.D.Includes bibliographical reference