Research Projects Directory

Scientific Approaches

We plan to search the database for people with COVID infection and/or vaccination and evaluate if there are any associations with dermatomyositis and scleroderma.

Scientific Approaches

Build ML models to discover the potentail patterns for the Covid 19 patients who had smell lose

Scientific Approaches

Build ML models to discover the potentail patterns for the Covid 19 patients who had smell lose

Scientific Approaches

To answer my scientific question, I plan to use a combination of epidemiological analysis and environmental data modeling. Specifically, I will use publicly available datasets on air quality, weather patterns, and reported cases of respiratory infections, such as flu and COVID-19, from health organizations and government sources. I will apply statistical methods, including regression analysis, to examine the relationships between environmental factors (like air pollution levels, temperature, and humidity) and the spread of infections over time. Additionally, I will use geographic information system (GIS) tools to visualize the spatial distribution of infections and environmental conditions across urban areas. This will help identify patterns and high-risk zones for outbreaks.

Scientific Approaches

To answer my scientific question, I plan to use a combination of epidemiological analysis and environmental data modeling. Specifically, I will use publicly available datasets on air quality, weather patterns, and reported cases of respiratory infections, such as flu and COVID-19, from health organizations and government sources. I will apply statistical methods, including regression analysis, to examine the relationships between environmental factors (like air pollution levels, temperature, and humidity) and the spread of infections over time. Additionally, I will use geographic information system (GIS) tools to visualize the spatial distribution of infections and environmental conditions across urban areas. This will help identify patterns and high-risk zones for outbreaks.

Scientific Approaches

To achieve this objective, data science workflows were used to apply ML algorithms on the Researcher Workbench. This effort allowed an expansion in the number of participants used to evaluate the ML models used to identify risk of PASC/Long COVID and also serve to validate the efforts of one team and providing insight to other teams. These models were implemented within the All of Us Controlled Tier data (C2022Q2R2), which was last refreshed on June 22, 2022. We intend to provide a step-by-step guide for the implementation of N3C's ML Model for identification of PASC/Long COVID Phenotype in the All of Us dataset.

Scientific Approaches

We will filter by patients diagnosed with Long COVID, identify trends in patient outcome with different treatments and other variables, and apply machine learning models to identify predictors of outcome using demographic and clinical data.

Scientific Approaches

We are using a retrospective cohort study design to examine factors associated with intubation in COVID-19 patients.
Dependent Variable: Binary variable indicating whether intubation occurred (1) or did not occur (0).
Independent Variables: Demographics: Age, sex, race, and ethnicity.
Physical Measurements: BMI and pregnancy status.
Biomarkers: D-dimer, Interleukin-6 (IL-6), IL-10, CRP. Collect biomarker data at specific times relative to the diagnosis of COVID-19.
COVID Vaccine Status: Document vaccination status and dates.
Drug: Record medications administered and their timing relative to COVID-19 diagnosis.

Repeated measures logistic regression will be performed to assess the relationship between the biomarkers and intubation status. Adjusted odds ratios (aOR) will be reported with their 95% confidence intervals. We will also examine potential interaction effects potential interactions between independent variables (e.g., biomarkers and demographics).

Scientific Approaches

we will use all of us research dataset, research method will be mostly cross sectional study, we will use the AI code generatior and our in house analyst to overview our statistical part

Scientific Approaches

This analysis will pull data from the Basics survey and the COPE surveys to examine whether answers in the COPE surveys can be broken down by differential housing circumstances. This will include summary and bivariate analyses.

Scientific Approaches

We will use the EHR, COPE survey, and in the future genetic information to explore associations between genetics and COVID as well as burden of disease.

Scientific Approaches

Using condition and medication information in the Controlled Tier dataset, we will look for associations between patients who used different diabetes medications prior to a COVID-19 infection to quantify their risk of developing PASC. An analytic fact table will be developed and data will be analyzed using Python and SQL. The study design is a retrospective cohort analysis using trial emulation techniques in adults with documented SARS-CoV-2 infection. The index date will be the date of first documented SARS-CoV-2 infection, and the exposure of interest: existing metformin or other diabetes medication prescription. The outcome of interest is a subsequent diagnosis of PASC.

Scientific Approaches

I plan to utilize available datasets related to COVID-19 vaccination, public health outcomes, and socioeconomic factors on All of Us platform.

Research Methods and Analytical Approaches:
(a) Exploratory Data Analysis (EDA)
Descriptive Statistics:
Compute means, medians, and standard deviations for vaccination rates, case fatality rates, and economic indicators.
Identify missing data and perform necessary cleaning (handling null values, duplicates).
Data Visualization:
Create time-series plots to observe vaccination progress vs. case and death rates.
Heatmaps to detect correlations between vaccination rates and socioeconomic factors.
(b) Statistical and Machine Learning Approaches
Correlation and Regression Analysis:
Pearson and Spearman correlation to assess relationships between vaccination rates, public health outcomes, and economic recovery.
Multiple Linear Regression to analyze the impact of vaccination rates on COVID-19 fatalities and economic growth.

Scientific Approaches

For each Long COVID disease signature from Taylor et al. (2023), we will generate summary statistics (e.g., # cases & controls, odds ratio, p-value) to evaluate the overall degree of replication in a patient cohort comprised of long COVID patients and healthy controls. Signatures with odds ratio <1 will be flagged as non-replicating. We will also test whether the count of disease signatures possessed by each patient is significantly associated with case-control status. This test will be repeated in ancestry-specific cohorts to identify potential challenges for health equity.
For each signature, we will evaluate the contribution of each component SNP to disease risk by comparing the odds ratio for patients with the full signature to the odds ratio for patients with the broader signature excluding the focal SNP. SNPs will be removed from the signature when the odds ratio of the latter exceeds the former. This refinement process will be repeated using a 5-fold cross validation approach.

Scientific Approaches

We will use data from the COPE surveys to assess suicidal ideation during the pandemic. We will link the external data capturing neighborhood social capital. We will then perform regression analysis, adjusting for many individual- and neighborhood-characteristics as potential confounders.

Scientific Approaches

Dataset: develop a dataset of Fitbit users with and without certain chronic conditions.
We will describe the sample in terms of sociodemographics. We will use the combination of feature engineering and machine learning techniques to assess differences between groups.

Scientific Approaches

Perform a retrospective analysis of COVID 19 and access the outcome as pulmonary hypertension. Would like to see if there is correlation between these diseases. If there is a correlation would like to explore the outcome across different gender, ethnicity and age based.

Scientific Approaches

Machine learning models will assess ICU risk using clinical data collected – allowing for early identification of high-risk patients. To improve fairness and accuracy, strategies will be applied to ensure the dataset is balanced, ensuring predictions remain reliable across different patient groups. The goal is to develop an interpretable and practical decision-support tool for ICU planning and resource allocation.

Mathematical component – within host models or transmission models – SIR-type models – Ordinary Differential based equations and analysis. – SARS-CoV-2 and Influenza seasonal variant.

Scientific Approaches

We will assemble a dataset of participants with and without cardiac arrhythmias and COVID-19 as well as relevant covariates. We will conduct logistic regression using both a matched case design and a case-crossover design in order to understand whether a significant relationship exists between COVID-19 infection and new onset cardiac arrythmia. Data cleaning will be conducted in python, and analysis will be conducted in R.

Scientific Approaches

The aim of my research is to determine the correlation between blood glucose, hemoglobin A1c and SARS-CoV-2 antibody among immunized type 2 diabetic patient who received different doses and types of COVID19 vaccines. I am planning on creating a dataset consisting of type 2 diabetic patients who received their COVID-19 vaccines and have their blood glucose, HbA1c, and SARS-CoV-2 antibody measurements available. I am planning on using SAS and various statistical methods to answer my research questions.

Scientific Approaches

Perform a retrospective analysis of COVID 19 and access the outcome as pulmonary hypertension. Would like to see if there is correlation between these diseases. If there is a correlation would like to explore the outcome across different gender, ethnicity and age based.

Scientific Approaches

A well-known challenge for machine learning in healthcare is that patients with different comorbidities induce different distributions over outcomes and the pattern changes across time. We will study two questions related to this heterogeneity, with the goal of producing more effective characterizations of patients at risk of severe outcomes due to SARS-COV-2 infection as well as more effective population-level tracking of COVID-19 outbreaks and disease burden. First, we will study how temporal variation impacts the risk factors and how to detect the shifts in distribution. Second, we will study how multiple sites could pool surveillance-related data together in a manner that preserves patient privacy and allows more rapid detection of new outbreaks or changes in the distribution of disease burden.

Scientific Approaches

PHQ-9 questions and answers will be retrieved for participants involved in six different time points. Response to each question will be converted to numeric scores (0, 1, 2, 3), and then summed up to derive the PHQ-9 total score. Participants missing any individual score were not included in this study. Binary status of suicidal ideation will be defined using item-9 answer (i.e., yes for >0). Distributions of PHQ-9 total score, suicidal ideation status at each time session will be reported by descriptive statistics stratified by age, sex, and ancestry. Their changes across different time sessions were tested by Kruskal-Wallis (KW) test, Friedman test, or chi-square test. Multivariable analyses are going to be conducted by generalized linear mixed models.

Scientific Approaches

Datasets: N3C is one of the richest data sources that include the electronic health records data for more than 5 million confirmed covid-19 patients from 74 sites
across the United States. All of Us is a unique source where we can access the genetic and clinical data for 100000 US patients and with higher representation for
minority groups.

Research Methods and Methods : We train a deep learning-based model on COVID-19 patients’ data available through the N3C initiative. As an outcome, the model will learn a phenotypic representation that consists of the patient's risk to develop post COVID complications, including neuropsychiatric complications Afterwards, we will transfer the model to the All of Us researcher platform and apply our model to create the phenotypic representation for the 11,767 COVID-19 patients using their EHR data. Then, we will use the genotypic data for 3,653 Covid-19 patients who has both their whole genome sequencing (WGS) and EHR data for the GWAS study

Scientific Approaches

We will create three data sets for analysis:

1. A list of diseases rated in the following ways:

a. Evidence Class
i. Strong evidence it is autoimmune
ii. Moderate evidence it is autoimmune
iii. Weak evidence for autoimmunity
iv. A comorbidity of autoimmune disease
v. Symptom or symptom set with no known mechanism

b. Autoinflammatory versus autoimmune flag

c. “Not always autoimmune” flag – to indicate diseases that could have alternative mechanisms of cause

2. A list of patients, anonymized, with socioeconomic, geographic and other data that would be of interest to patients and public health officials to understand which communities are affected by these diseases
3. Outcomes data for patients over time assessing quality of life using PROMIS metrics