Skip to main content
eScholarship
Open Access Publications from the University of California

Scholarly Works (18 results)

  • Thesis
  • Peer Reviewed
UCLA Electronic Theses and Dissertations (2022)

Inferring dependencies between complex biological traits while accounting for evolutionary relationships among specimens is of great scientific interest, yet remains infeasible when trait and specimen counts grow large. I aim to develop a scalable Bayesian inference framework to assess correlation between complex traits along the evolutionary tree relating the specimens and informed by molecular sequences. To accommodate discrete and continuous traits, I posit a phylogenetic multivariate probit model that uses a latent variable framework. Posterior computation under this model requires integrating many latent variables, or equivalently making many computationally expensive draws from a high-dimensional multivariate truncated normal distribution (MTN). To tackle this challenge, I propose an inference scheme that exploits 1) representative cutting-edge Markov chain Monte Carlo (MCMC) methods including the bouncy particle sampler (BPS), the Markovian Zigzag sampler (ZZ), and the Zigzag Hamiltonian Monte Carlo (Zigzag-HMC) that can simultaneously sample all truncated normal dimensions, and 2) novel dynamic programming strategies that reduce the cost of likelihood and gradient evaluations for all three samplers to linear in sample size. Compared to the previous best practices that employ multiple-try rejection sampling, my approach achieves an order-of-magnitude speedup, allowing us to tackle previously unworkable large-scale problems. In an application with 535 HIV-1 viruses and 24 traits that necessitates sampling from a 11,235-dimensional MTN, my method makes it possible to examine the conditional dependencies between 21 immune escape mutations and 3 virulence measurements. In a second application I study the evolution of influenza H1N1 glycosylations using around 900 viruses. Lastly, I extend the phylogenetic probit model to incorporate categorical traits and demonstrate its use to investigate Aquilegia flower and pollinator coevolution. In summary, the contribution of this dissertation is two-fold. First, I develop a state-of-the-art solution for the long-standing problem in Bayesian phylogenetics | learning correlation among complex biological traits with joint tree modeling. Second, further empirical and theoretical investigation of BPS, ZZ, and Zigzag-HMC yield insight into the differences and similarities between these recently developed MCMC samplers. As Zigzag-HMC outperforms the other two on MTNs, I also implement this approach in a standalone R package, aiming to provide a general efficient tool for high-dimensional MTN simulation.

    Cover page: Large-scale Inference of Correlation between Complex Biological Traits
    • Thesis
    • Peer Reviewed
    UC Irvine Electronic Theses and Dissertations (2022)

    Although automation has become widespread in many industries, some workplaces, such as utility industries, still rely heavily on individuals to perform critical tasks based on their extensive past experiences. This accumulated intellectual capital not only improves productivity and efficacy under normal conditions but more importantly, also contains knowledge about identifying anomalies and addressing unexpected events and situations. This knowledge is essential and critical to ensuring system safety and reliability, especially for industries with aging infrastructures where anomalies are becoming more common. Unfortunately, this knowledge is not explicitly recorded in defined guidelines, protocols, or standard workflows but implicitly resides in the minds of skilled workers and routine event logs. As skilled workers retire or leave the business, we may lose this accrued indispensable knowledge that is critical to the industries’ productivity, reliability, and safety.This thesis proposes a framework to discover implicit knowledge from event logs. The implementation of the framework contains three phases. We first proposed an approach that uses hidden Markov models to filter out outliers from event logs in the clean phase. Then, in the discovery phase, we propose approaches to discover the implicit knowledge related to timing constraints and process scenarios from event logs. For timing constraint discovery, we have presented an approach that extends existing process mining techniques to mine and integrates timing constraints with a workflow or process model constructed by any existing process mining algorithm. A real-life road traffic fine management process scenario is used as a case study to investigate the effectiveness and validity of the approach. For process scenario discovery, we present the distance-based and density-based approaches that obtain timing information from event logs and use the information to assist process scenario discoveries. A wastewater treatment process provided by a domain expert is used as a case study to investigate the effectiveness and validity of the approach. In incorporating phase, we present an approach that incorporates domain knowledge to assist in discovering process scenarios from event logs. The case studies and experiment results indicate that the proposed implicit knowledge discovery framework can mine implicit knowledge related to the timing constraints and process scenarios from event logs and outperform commonly used approaches with different real-life event logs.

      Cover page: A Framework for Discovering Implicit Knowledge from Event Logs
      • Article
      • Peer Reviewed
      UCLA Previously Published Works (2023)
      Inferring dependencies between mixed-type biological traits while accounting for evolutionary relationships between specimens is of great scientific interest yet remains infeasible when trait and specimen counts grow large. The state-of-the-art approach uses a phylogenetic multivariate probit model to accommodate binary and continuous traits via a latent variable framework, and utilizes an efficient bouncy particle sampler (BPS) to tackle the computational bottleneck-integrating many latent variables from a high-dimensional truncated normal distribution. This approach breaks down as the number of specimens grows and fails to reliably characterize conditional dependencies between traits. Here, we propose an inference pipeline for phylogenetic probit models that greatly outperforms BPS. The novelty lies in 1) a combination of the recent Zigzag Hamiltonian Monte Carlo (Zigzag-HMC) with linear-time gradient evaluations and 2) a joint sampling scheme for highly correlated latent variables and correlation matrix elements. In an application exploring HIV-1 evolution from 535 viruses, the inference requires joint sampling from an 11,235-dimensional truncated normal and a 24-dimensional covariance matrix. Our method yields a 5-fold speedup compared to BPS and makes it possible to learn partial correlations between candidate viral mutations and virulence. Computational speedup now enables us to tackle even larger problems: we study the evolution of influenza H1N1 glycosylations on around 900 viruses. For broader applicability, we extend the phylogenetic probit model to incorporate categorical traits, and demonstrate its use to study Aquilegia flower and pollinator co-evolution.
        Cover page: Accelerating Bayesian inference of dependency between mixed-type biological traits.
        • Article
        • Peer Reviewed
        UCLA Previously Published Works (2021)
        Relaxed random walk (RRW) models of trait evolution introduce branch-specific rate multipliers to modulate the variance of a standard Brownian diffusion process along a phylogeny and more accurately model overdispersed biological data. Increased taxonomic sampling challenges inference under RRWs as the number of unknown parameters grows with the number of taxa. To solve this problem, we present a scalable method to efficiently fit RRWs and infer this branch-specific variation in a Bayesian framework. We develop a Hamiltonian Monte Carlo (HMC) sampler to approximate the high-dimensional, correlated posterior that exploits a closed-form evaluation of the gradient of the trait data log-likelihood with respect to all branch-rate multipliers simultaneously. Our gradient calculation achieves computational complexity that scales only linearly with the number of taxa under study. We compare the efficiency of our HMC sampler to the previously standard univariable Metropolis-Hastings approach while studying the spatial emergence of the West Nile virus in North America in the early 2000s. Our method achieves at least a 6-fold speed increase over the univariable approach. Additionally, we demonstrate the scalability of our method by applying the RRW to study the correlation between five mammalian life history traits in a phylogenetic tree with $3650$ tips.[Bayesian inference; BEAST; Hamiltonian Monte Carlo; life history; phylodynamics, relaxed random walk.].
          Cover page: Relaxed Random Walks at Scale
          • Article
          • Peer Reviewed
          UCLA Previously Published Works (2020)
          Calculation of the log-likelihood stands as the computational bottleneck for many statistical phylogenetic algorithms. Even worse is its gradient evaluation, often used to target regions of high probability. Order O(N)-dimensional gradient calculations based on the standard pruning algorithm require O(N2) operations, where N is the number of sampled molecular sequences. With the advent of high-throughput sequencing, recent phylogenetic studies have analyzed hundreds to thousands of sequences, with an apparent trend toward even larger data sets as a result of advancing technology. Such large-scale analyses challenge phylogenetic reconstruction by requiring inference on larger sets of process parameters to model the increasing data heterogeneity. To make these analyses tractable, we present a linear-time algorithm for O(N)-dimensional gradient evaluation and apply it to general continuous-time Markov processes of sequence substitution on a phylogenetic tree without a need to assume either stationarity or reversibility. We apply this approach to learn the branch-specific evolutionary rates of three pathogenic viruses: West Nile virus, Dengue virus, and Lassa virus. Our proposed algorithm significantly improves inference efficiency with a 126- to 234-fold increase in maximum-likelihood optimization and a 16- to 33-fold computational performance increase in a Bayesian framework.
            Cover page: Gradients Do Grow on Trees: A Linear-Time O(N)-Dimensional Gradient for Statistical Phylogenetics.
            • Article
            • Peer Reviewed
            UCLA Previously Published Works (2023)
            Researchers studying the evolution of viral pathogens and other organisms increasingly encounter and use large and complex data sets from multiple different sources. Statistical research in Bayesian phylogenetics has risen to this challenge. Researchers use phylogenetics not only to reconstruct the evolutionary history of a group of organisms, but also to understand the processes that guide its evolution and spread through space and time. To this end, it is now the norm to integrate numerous sources of data. For example, epidemiologists studying the spread of a virus through a region incorporate data including genetic sequences (e.g. DNA), time, location (both continuous and discrete) and environmental covariates (e.g. social connectivity between regions) into a coherent statistical model. Evolutionary biologists routinely do the same with genetic sequences, location, time, fossil and modern phenotypes, and ecological covariates. These complex, hierarchical models readily accommodate both discrete and continuous data and have enormous combined discrete/continuous parameter spaces including, at a minimum, phylogenetic tree topologies and branch lengths. The increased size and complexity of these statistical models have spurred advances in computational methods to make them tractable. We discuss both the modeling and computational advances below, as well as unsolved problems and areas of active research.
              Cover page: Data Integration in Bayesian Phylogenetics
              • Article
              • Peer Reviewed
              UC Davis Previously Published Works (2020)
              Dietary intake of linoleic acid (LA, 18:2ω-6) has risen dramatically in recent decades. Previous studies have suggested a high intake of LA could increase tissue concentrations of proinflammatory and protumorigenic ω-6-series eicosanoid metabolites, increasing risks of inflammation and associated diseases. However, the effects of a LA-rich diet on in vivo profiles of eicosanoids and development of inflammatory diseases are understudied. Here, we treated spontaneous colitis-prone (Il-10-/-) mice with a control diet (∼3 Cal% LA) or a LA-rich diet (∼9 Cal% LA) for 18 weeks and analyzed the effects of the LA-rich diet on profiles of eicosanoids and development of colitis. We found that treatment with the LA-rich diet increased the tissue level of LA: the liver levels of LA were 5.8 ± 0.6% in the control diet-treated mice versus 11.7 ± 0.7% in the LA-rich diet-treated mice (P < 0.01). The plasma concentrations of a series of LA-derived metabolites, including 9-hydroxyoctadecadienoic acid (HODE), 9,10-dihydroxyoctadecenoic acid (DiHOME), 12,13-DiHOME, and 13-HODE were significantly increased by treatment with the LA-rich diet (P < 0.05). However, the LA-rich diet had little effect on the severity of colitis in the treated Il-10-/- mice. These results suggest a limited role of increased consumption of dietary LA on promoting colitis in the Il-10-/- model.
                Cover page: Effects of Linoleic Acid-Rich Diet on Plasma Profiles of Eicosanoids and Development of Colitis in Il-10–/– Mice
                • Article
                • Peer Reviewed
                UC Riverside Previously Published Works (2015)

                Background and purpose

                Proteinase-activated receptor-2 (PAR2) is a GPCR linked to diverse pathologies, including acute and chronic pain. PAR2 is one of the four PARs that are activated by proteolytic cleavage of the extracellular amino terminus, resulting in an exposed, tethered peptide agonist. Several peptide and peptidomimetic agonists, with high potency and efficacy, have been developed to probe the functions of PAR2, in vitro and in vivo. However, few similarly potent and effective antagonists have been described.

                Experimental approach

                We modified the peptidomimetic PAR2 agonist, 2-furoyl-LIGRLO-NH2 , to create a novel PAR2 peptidomimetic ligand, C391. C391 was evaluated for PAR2 agonist/antagonist activity to PAR2 across Gq signalling pathways using the naturally expressing PAR2 cell line 16HBE14o-. For antagonist studies, a highly potent and specific peptidomimetic agonist (2-aminothiazo-4-yl-LIGRL-NH2 ) and proteinase agonist (trypsin) were used to activate PAR2. C391 was also evaluated in vivo for reduction of thermal hyperalgesia, mediated by mast cell degranulation, in mice.

                Key results

                C391 is a potent and specific peptidomimetic antagonist, blocking multiple signalling pathways (Gq -dependent Ca2+ , MAPK) induced following peptidomimetic or proteinase activation of human PAR2. In a PAR2-dependent behavioural assay in mice, C391 dose-dependently (75 μg maximum effect) blocked the thermal hyperalgesia, mediated by mast cell degranulation.

                Conclusions and implications

                C391 is the first low MW antagonist to block both PAR2 Ca2+ and MAPK signalling pathways activated by peptidomimetics and/or proteinase activation. C391 represents a new molecular structure for PAR2 antagonism and can serve as a basis for further development for this important therapeutic target.
                  Cover page: PAR2 antagonist C391
                  • Article
                  • Peer Reviewed
                  UCLA Previously Published Works (2023)
                  Human group B Streptococcus (GBS) infections attributable to an invasive, hypervirulent sequence type (ST) 283 have been associated with freshwater fish consumption in Asia. The origin, geographic dispersion pathways and host transitions of GBS ST283 remain unresolved. We gather 328 ST283 isolate whole-genome sequences collected from humans and fish between 1998 and 2021, representing eleven countries across four continents. We apply Bayesian phylogeographic analyses to reconstruct the dispersal history of ST283 and combine ST283 phylogenies with genetic markers and host association to investigate host switching and the gain and loss of antimicrobial resistance and virulence factor genes. Initial dispersal within Asia followed ST283 emergence in the early 1980s, with Singapore, Thailand and Hong Kong observed as early transmission hubs. Subsequent intercontinental dispersal originating from Vietnam began in the decade commencing 2001, demonstrating ST283 holds potential to expand geographically. Furthermore, we observe bidirectional host switching, with the detection of more frequent human-to-fish than fish-to-human transitions, suggesting that sound wastewater management, hygiene and sanitation may help to interrupt chains of transmission between hosts. We also show that antimicrobial resistance and virulence factor genes were lost more frequently than gained across the evolutionary history of ST283. Our findings highlight the need for enhanced surveillance, clinical awareness, and targeted risk mitigation to limit transmission and reduce the impact of an emerging pathogen associated with a high-growth aquaculture industry.
                    Cover page: Dispersal history and bidirectional human-fish host switching of invasive, hypervirulent Streptococcus agalactiae sequence type 283.
                    • Article
                    • Peer Reviewed
                    UCLA Previously Published Works (2019)

                    Background

                    High-sensitivity cardiac troponin (hs-cTn) elevations are associated with incident cardiovascular disease events in primary prevention samples. However, the mechanisms underlying this association remain unclear.

                    Methods

                    We studied 458 men without known cardiovascular disease who participated in the cardiovascular disease substudy of the Multicenter AIDS Cohort Study and had cardiac CT angiography. We used multivariable linear and logistic regression models to examine the cross-sectional associations between coronary artery stenosis, coronary artery plaque, indexed left ventricular mass (LVMi), and the outcome of hs-cTnI. We also evaluated the associations between HIV serostatus or use of highly active antiretroviral therapy (HAART) and hs-cTnI.

                    Results

                    The mean age was 54 years, 54% were white, and 61% were HIV infected. In multivariable-adjusted logistic models, comparing the highest quartile of LVMi with the lowest quartile, the odds ratio (OR) of hs-cTnI ≥75th percentile was 2.59 (95% CI, 1.20-5.75). There was no significant association between coronary stenosis severity or plaque type and hs-cTnI in linear models; however, in logistic regression models, coronary artery stenosis ≥70% (8% of sample) was marginally associated with a higher likelihood (OR, 2.75 [95% CI, 1.03, 7.27]) of having hs-cTnI ≥75th percentile. There were no associations between HIV serostatus or HAART use and hs-cTnI in either linear or logistic models.

                    Conclusion

                    Among primary prevention men with or at risk for HIV, hs-cTnI concentrations were strongly associated with LVMi but were not associated with HIV infection or treatment status or with coronary plaque type or stenosis until the extremes of severity (≥70% stenosis).
                      Cover page: Association of High-Sensitivity Troponin with Cardiac CT Angiography Evidence of Myocardial and Coronary Disease in a Primary Prevention Cohort of Men: Results from MACS
                      Top