Trapped radioactive atoms are an appealing source for precise measurements of the beta-neutrino correlation coefficient, a, since the momentum of the neutrino can be inferred from the detection of the unperturbed low-energy recoil daughter nucleus. Sodium-21 is produced on-line at the 88" cyclotron at Lawrence Berkeley National Laboratory, and 8e5 atoms have been maintained in a magneto-optical trap. A static electric field draws daughter Neon-21 ions to a microchannel plate detector and betas are detected in coincidence with a plastic scintillator beta detector. The Neon-21 time-of-flight distribution determines the beta neutrino correlation coefficient, a. The resulting charge-state distribution is compared to a simple model based on the sudden approximation which suggests a small but important contribution from nuclear recoil-induced ionization. A larger than expected fraction of the daughters are detected in positive charge-states, but no dependence on either the beta or recoil nucleus energy was observed. We find a = 0.5243 plus or minus 0.0092, which is in 3.6 sigma disagreement with the Standard Model prediction of a = 0.559 plus or minus 0.003. Aside from a deviation from the Standard Model, a possible explanation for the discrepancy is that the branching ratio to the first excited state is in error.

While prior research has found racial disparities in the administration of death sentences, less is known about the processes generating these patterns. To understand how racial disparities are formed and sustained within death penalty institutions, this study tracks homicide cases as they pass through multiple stages of Los Angeles County’s criminal justice system. Drawing upon the notion of cumulative disadvantage—a process by which initial disadvantages in group-positionality lead to additional relative loses overtime—I focus on the accumulation of racial biases across multiple decision-making points. In chapter 1, multi-level logistic regressions disentangle the effects of agency, neighborhood, and case characteristics on homicide arrests. While several non-racial factors influence arrest patterns, homicides involving minority victims and those occurring in neighborhoods with large minority populations are less likely to be solved. Chapter 2 expands upon these insights, exploring the link between homicide arrests and charging practices. Two-stage selection models indicate that cases involving minority victims are less likely to contain a death penalty eligible charge, and these effects are mediated by the likelihood of arrest. Although defendant race is less influential, it moderates victim race effects, such that Black-on-White homicides are more likely to receive a death penalty eligible charge. Chapter 3 integrates the first two studies by investigating the accumulation of racial disparities across multiple stages of the criminal justice system. Ordered-logistic regressions show that cases involving minority victims are less likely to advance to capital trial, in part, because of racial disparities at earlier stages in the process. Defendant race effects are less consistent, but frequently condition the influence of victim race, with cases involving White victims and minority defendants receiving harsher punishment. Taken together, the dissertation suggests that racial disparities within capital punishment systems arise from a complex chain of decisions, rather than any single decision-making point. These patterns speak to the institutional role of race within criminal justice systems, offering support for the cumulative disadvantage perspective. Moreover, results contribute to contemporary policy debates by highlighting the need for multi-stage policy reforms.

This research focuses on three primary topics: (1) the development of biolayer interferometry serology tests to detect viral antigen reactivity with microliter scale human serum samples; (2) characterization of targets and epitopes for antibody-based cancer therapeutics with an emphasis on immune checkpoint regulation of natural killer cells via KIR receptors; and (3) preliminary investigation of the human astrovirus 1 RNA-dependent-RNA-polymerase.

BLI-ISA: The biolayer interferometry based antibody test (BLI-ISA) represents a novel, high throughput platform with numerous advantages over conventional enzyme-linked immunosorbent assays (ELISA) in terms of sample requirements, labor, and overall assay time while maintaining high accuracy. Compared to the COV2T antibody test for SARS-CoV-2, BLI-ISA had 94\% positive percent agreement and >98\% negative percent agreement in detection of serum IgG reactivity to viral antigen. The BLI-ISA antibody test is readily adaptable to new antigens, and was quickly deployed to assess reactivity to a panel of eight human astrovirus serotypes with good agreement to studies using more laborious methods.

Antibody therapeutics for cancer: a comparison of bioinformatic methods for the identification of tumor surface markers using RNAseq reveals statistical methods like DESeq outperform some dimensionality reduction techniques like PCA. Then, the epitope of an immune checkpoint inhibitor antibody is characterized by X-ray crystallography in the context of KIR receptor variants, highlighting the importance of population-wide polymorphism analysis.

HAstV RdRp: expression methods for this under-studied polymerase were developed to aid future structural work and activity assays.