This project initially aimed to recreate the following analysis associating long-term air pollution exposure with lung cancer risk: Huang et al., “Air Pollution, Genetic Factors, and the Risk of Lung Cancer.” Replication of results was successful.
These analyses were extended to review the association between air pollution exposure and the risk of developing clonal hematopoiesis.
- Open data_process.R, change prefix path to match username, save
- Open master.Rmd in RStudio
- Click the dropdown on Knit, select "Knit with Parameters"
- Select data to use and whether to include code in the output
- Once the report has been generated, review the temporary HTML
- Make bmiCat factor variable (Normal (<25 kg/m2), Overweight (25–29.9 kg/m2), Obesity (>=30 kg/m2, Missing value)
- Find average pack-years of smoking
- Simplify education levels. Include each participant in only one level. (Degree level or professional education, Other levels, Missing value)
- Simplify householdIncomeCat (Less than £31,000, £31,000 and above, Missing value)
- Simplify smokingCat (Never-smoker, Current or former smoker, Missing value)
- Make categorical variables into factors and set reference level
- Simplify fuel (yes/no gas or solid-fuel cooking/heating in home, Missing value)
- Make exposeSmokeHome factor variable. Remove raw variable from the model
- histology variable
- Include genetic component variables
- Table of population characteristics using table1
- Incidence rate of phenotype
- Number of subjects excluded during filtering
- Distributions of all relevant variables
- Pearson correlation coefficients between pollutants
- Pearson correlation coefficients between covariates
- Data processing for CH (CH missing if exome failed. CH-PD = CH=1 & PD=1)
- Update CH file when the latest results become available (200k participants currently)
- Test running on a cluster. Update file paths as needed
- Lung PRS - ask Xiaoyu for this data
- Check for violation of proportional hazard (constant HR over time)
- Model all pollutants, including genetic components
- Tabs for navigating between pollutants
- Stratified analyses (Sex, Age, Obesity, Smoking Status, Household Income, Histology)
- Consider whether common skin cancers should be included in prior cancer filtering
- Explore Genetic Analysis of Time-to-Event phenotypes (GATE) Method to adjust for population structure, sample relatedness, and p-value inflation due to very low-frequency variants and heavily censored phenotypes
- Replace dense scatter plots with hexplots
- Generate exposeSmokeHomeImputed using mice package
- Sensitivity Analyses
- Host data using Github Large Files