A5076555314- Statistical Methods and Inference
- Statistical Methods and Bayesian Inference
- Bayesian Methods and Mixture Models
- Advanced Causal Inference Techniques
- Anomaly Detection Techniques and Applications
- Network Security and Intrusion Detection
- Smoking Behavior and Cessation
- Insurance, Mortality, Demography, Risk Management
- Advanced Clustering Algorithms Research
- HIV Research and Treatment
- Genetic Associations and Epidemiology
- Genetic Mapping and Diversity in Plants and Animals
- Point processes and geometric inequalities
- HIV/AIDS Research and Interventions
- Natural Language Processing Techniques
- Risk and Safety Analysis
- Health Promotion and Cardiovascular Prevention
- Dementia and Cognitive Impairment Research
- Cardiovascular Health and Risk Factors
- Statistical Distribution Estimation and Applications
- Animal Virus Infections Studies
- Text and Document Classification Technologies
- Gene expression and cancer classification
- Random Matrices and Applications
- Linguistics and Discourse Analysis
University of Manitoba
2021-2025
McGill University
2015-2019
McGill University Health Centre
2019
Institute for Medical Research
2016
Jewish General Hospital
2015-2016
Abstract In observational studies, type-2 diabetes (T2D) is associated with an increased risk of coronary heart disease (CHD), yet interventional trials have shown no clear effect glucose-lowering on CHD. Confounding may therefore influenced these estimates. Here we use Mendelian randomization to obtain unconfounded estimates the influence T2D and fasting glucose (FG) CHD risk. Using multiple genetic variants FG, find that increases (odds ratio (OR)=1.11 (1.05–1.17), per unit increase in...
Microbial translocation from the gut to systemic circulation contributes immune activation during human immunodeficiency virus (HIV) infection and is usually assessed by measuring plasma levels of bacterial lipopolysaccharide (LPS). Fungal colonization in increases HIV-infection people living with HIV (PLWH) have increased fungal polysaccharide (1→3)-β-D-Glucan (βDG). We contribution circulating DG PLWH.Cross-sectional longitudinal assessments βDG were conducted along markers disease...
Regenerating islet-derived protein 3α (REG3α) is an antimicrobial peptide secreted by intestinal Paneth cells. Circulating REG3α has been identified as a gut damage marker in inflammatory bowel diseases. People living with human immunodeficiency virus (PWH) on antiretroviral therapy (ART) present abnormal landscape leading to microbial translocation, persistent inflammation, and development of non-AIDS comorbidities. Herein, we assessed PWH.
ABSTRACT Background Older persons living with dementia (PLWD) often have multiple other chronic health conditions (i.e., comorbidities). Network analyses can describe complex profiles of through graphical displays grounded in empirical data. Our study compared patterns among PLWD residing and outside long‐term care (LTC) settings. Methods Population‐based administrative data, including outpatient physician claims, inpatient records, pharmaceutical LTC for the were from Canadian province...
An unprecedented amount of SARS-CoV-2 sequencing has been performed, however, novel bioinformatic tools to cope with and process these large datasets is needed. Here, we have devised a pipeline that inputs genome in FASTA/FASTQ format outputs single Variant Calling Format file can be processed obtain variant annotations perform downstream population genetic testing. As proof concept, analyzed over 229,000 viral sequences up until November 30, 2020. We identified 39,000 variants worldwide...
Unsupervised anomaly detection (UAD) is a diverse research area explored across various application domains. Over time, numerous techniques, including clustering, generative, and variational inference-based methods, are developed to address specific drawbacks advance state-of-the-art techniques. Deep learning generative models recently played significant role in identifying unique challenges devising advanced approaches. Auto-encoders (AEs) represent one such powerful technique that combines...
Unsupervised Anomaly Detection (UAD) is a highly diverse research area explored across various application domains. Throughout the years, numerous anomaly detection techniques, including clustering, generative, and variational inference-based methods, have been developed to address specific deficiencies enhance state-of-the-art approaches. Notably, deep learning generative models played crucial role in identifying unique problems introducing advanced approaches recent times.Among these,...
Clustering techniques are used to group observations and discover interesting patterns within data. Model-based clustering is one such method that often an attractive choice due the specification of a generative model for given data ability calculate model-selection criteria, which in turn select number clusters. However, when only distances between available, model-based can no longer be used, heuristic algorithms without aforementioned advantages usually instead. As solution, Oh Raftery...
In this paper, we consider some potential pitfalls of the growing use quasi‐likelihood‐based information criteria for longitudinal data to select a working correlation structure in generalized estimating equation framework. particular, examine settings where fully conditional mean does not equal marginal as well hypothesis testing following selection matrix. Our results suggest that any criterion matrix is inappropriate when model assumption violated. We also find type I error differs from...
The genomics era has led to an increase in the dimensionality of data collected investigation biological questions. In this context, dimension-reduction techniques can be used summarise high-dimensional signals into low-dimensional ones, further test for association with one or more covariates interest. This paper revisits such approach, previously known as principal component heritability and renamed here explained variance (PCEV). As its name suggests, PCEV seeks a linear combination...
In clinical studies of time-to-event data, a quantity interest to the clinician is their patient's risk an event. However, methods relying on time matching or risk-set sampling (including Cox regression) eliminate baseline hazard from estimating function. As consequence, focus has been reporting ratios instead survival cumulative incidence curves. Indeed, patient requires separate estimation hazard. Using case-base sampling, Hanley & Miettinen (2009) explained how parametric functions can be...
Dimension reduction techniques are among the most essential analytical tools in analysis of high-dimensional data. Generalized principal component (PCA) is an extension to standard PCA that has been widely used identify low-dimensional features discrete data, such as binary, multi-category and count For microbiome data particular, multinomial a natural counterpart PCA. However, this technique fails account for excessive number zero values, which frequently observed To allow sparsity,...
Objectives We developed a machine-learning model-based algorithm (MBA) for smoking in Administrative Health Data (AHD). The validity of this MBA was compared to rule-based (RBA). Approaches study included adults (≥18 years) from clinical registry containing self-reported current 2017 2020 Manitoba, Canada. Clinical data were linked up five years hospitalization, physician billing claims, and prescription medication records. RBA based on diagnosis codes tobacco use nicotine dependence...
Abstract The genomics era has led to an increase in the dimensionality of data collected investigate biological questions. In this context, dimension-reduction techniques can be used summarize high-dimensional signals into low-dimensional ones, further test for association with one or more covariates interest. This paper revisits such approach, previously known as Principal Component Heritability and renamed here Explained Variance (PCEV). As its name suggests, PCEV seeks a linear...
In epidemiological studies of time-to-event data, a quantity interest to the clinician and patient is risk an event given covariate profile. However, methods relying on time matching or risk-set sampling (including Cox regression) eliminate baseline hazard from likelihood expression estimating function. The then needs be estimated separately using non-parametric approach. This leads step-wise estimates cumulative incidence that are difficult interpret. Using case-base sampling, Hanley &...
Deep learning (DL) based natural language processing (NLP) has recently grown as one the fastest research domain and retained remarkable improvement in many applications. Due to significant amount of data, adaptation feature symmetric data efficiency is a critical underlying task such However, their ability extract features limited due lack proper model formation. Moreover, use these methods on smaller datasets unexplored underdeveloped compared more popular areas. This work introduces...
Clustering techniques are used to group observations and discover interesting patterns within data. Model-based clustering is one such method that often an attractive choice due the specification of a generative model for given data ability calculate model-selection criteria, which in turn select number clusters. However, when only distances between available, model-based can no longer be used, heuristic algorithms without aforementioned advantages usually instead. As solution, Oh Raftery...