initial commit

.Rprofile

@@ -0,0 +1 @@
+source("renv/activate.R")

.gitignore

@@ -0,0 +1,55 @@
+# History files
+.Rhistory
+.Rapp.history
+
+# Session Data files
+.RData
+.RDataTmp
+
+# User-specific files
+.Ruserdata
+
+# Example code in package build process
+*-Ex.R
+
+# Output files from R CMD build
+/*.tar.gz
+
+# Output files from R CMD check
+/*.Rcheck/
+
+# RStudio files
+.Rproj.user/
+
+# produced vignettes
+vignettes/*.html
+vignettes/*.pdf
+
+# OAuth2 token, see https://github.com/hadley/httr/releases/tag/v0.3
+.httr-oauth
+
+# knitr and R markdown default cache directories
+*_cache/
+/cache/
+
+# Temporary files created by R markdown
+*.utf8.md
+*.knit.md
+
+# R Environment Variables
+.Renviron
+
+# pkgdown site
+docs/
+
+# translation temp files
+po/*~
+
+# Large files
+results/largefiles/
+
+# Temporary word files
+*$*.docx
+
+# Ignore the supplement cause quarto is making the rendered doc huge
+products/manuscript/supplement.docx

Billings-2024-HD-Heterologous.Rproj

@@ -0,0 +1,13 @@
+Version: 1.0
+
+RestoreWorkspace: Default
+SaveWorkspace: Default
+AlwaysSaveHistory: Default
+
+EnableCodeIndexing: Yes
+UseSpacesForTab: No
+NumSpacesForTab: 2
+Encoding: UTF-8
+
+RnwWeave: Sweave
+LaTeX: pdfLaTeX

README.md

@@ -0,0 +1,24 @@
+# High dose inactivated influenza vaccine inconsistently improves heterologous antibody responses in an elderly human cohort
+
+This repo contains the code, data, and results associated with the manuscript
+"High dose inactivated influenza vaccine inconsistently improves heterologous antibody responses in an elderly human cohort"
+by Billings et al.
+
+## Reproducability instructions
+
+* You can find the reproducibility instructions in [the Supplement](./products/manuscript/supplement.pdf).
+* This code was generated with R 4.4.1 and the package versions specified in
+`renv`. If you don't install R 4.4.1 and use `renv::restore()` to get the
+correct package versions, it might not work.
+
+## Licensing
+
+All code written by us is licensed under the [GNU Affero General Public License,
+version 3](./LICENSE.md).
+
+## Contact information
+
+Please contact Zane Billings (wesley dot billings at uga dot edu) or
+Andreas Handel (ahandel at uga dot edu) for more information.
+
+<!-- end of file -->

code/01-Data-Processing.R

@@ -0,0 +1,141 @@
+###
+# Data Processing
+# Zane Billings
+# 2024-01-29
+# Starting with the cleaned data from ahgroup/UGAFluVac repo, we need to do
+# some data processing to get the data ready for our analyses.
+# This script processes and filters the data.
+###
+
+# Setup ========================================================================
+# Declare necessary dependencies
+library(readr, include.only = NULL)
+library(here, include.only = NULL)
+library(dplyr, include.only = NULL)
+library(forcats, include.only = NULL)
+
+# Load the "raw" data
+# This data is the output of the cleaning process from the repo
+# https://github.com/ahgroup/UGAFluVac-data/, which includes the raw excel
+# data files. We pulled the cleaned data from that repo on 2024-01-29.
+raw_data <- readr::read_rds(here::here("data", "raw", "clean-data.Rds"))
+
+# Data cleaning ================================================================
+# First filter out the records we don't want
+dat_filtered <-
+	raw_data |>
+	dplyr::filter(
+		# Remove individuals with intradermal vaccines
+		dose %in% c("SD", "HD"),
+		# Remove individuals with infinite titerincrease (1 person who has a
+		# pretiter of 0, probably a data entry error)
+		is.finite(titerincrease),
+		# Remove flu B assays
+		strain_type %in% c("H1N1", "H3N2"),
+		# Only individuals >= age 65 were allowed to get the HD vaccine, so we
+		# want to filter out any records from individuals below that age.
+		age >= 65
+	) |>
+	# Drop the missing factor levels that remain after filtering
+	dplyr::mutate(
+		# Change the sichuan 1989 (incorrect) records to 1987 so they are all
+		# listed as the same strain, which they are, due to the insane data
+		# cleaning error. SHould eventually be moved to the UGAFluVac repo.
+		strains_fullname = forcats::fct_recode(
+			strains_fullname,
+			"H3N2-Sichuan-1987" = "H3N2-Sichuan-1989"
+		),
+		# Now drop the factor levels.
+		dplyr::across(
+			c(dose, strain_type, strains_fullname),
+			forcats::fct_drop
+		)
+	)
+
+# Now we need to make edits to certain variables and drop the variables that
+# we don't need
+dat_clean <-
+	dat_filtered |>
+	# First we need to "condense" the vaccine name variables. We just need one
+	# vaccine name variable that has the name of the H1 vaccine strain for H1
+	# assays, and the name of the H3 vaccine strain for H3 assays.
+	dplyr::mutate(
+		vaccine_name = dplyr::if_else(
+			strain_type == "H1N1",
+			h1n1_vaccine_fullname,
+			h3n2_vaccine_fullname
+		),
+		.after = dose
+	) |>
+	# Now drop the *_vaccine_fullname columns -- we'll also drop out the
+	# variables we don't plan to use at all in this analysis.
+	dplyr::select(
+		-dplyr::ends_with("vaccine_fullname"),
+		-c(uniq_id, id, bmi, days_since_vac, date_vaccinated, race, gender),
+		-dplyr::contains("particip")
+	) |>
+	# Next we need to clean up the date of birth variable. We just need a variable
+	# for the birth year.
+	dplyr::mutate(
+		birth_year = substring(dateofbirth, 1, 4) |> as.integer(),
+		.keep = "unused"
+	) |>
+	# Now we'll create a fold change variable which is 2 ^ (titerincrease)
+	# which can make plotting easier
+	dplyr::mutate(fold_change = 2 ^ titerincrease, .before = titerincrease) |>
+	# Next let's transform the season variable into an ordered factor to ensure
+	# it behaves the way we want during plotting. Since the levels are already
+	# alphabetical it should be fine but better to go ahead and do it.
+	dplyr::mutate(season = factor(season, ordered = TRUE)) |>
+	# Finally let's clean up the names of the some of the variables.
+	dplyr::rename(
+		strain_name = strains_fullname,
+		posttiter = postiter,
+		log_pretiter = prevactiter,
+		log_posttiter = postvactiter,
+		titer_increase = titerincrease,
+		id = subject_id
+	)
+
+# Now we need to make a few additional changes before we can pass the
+# data to our models.
+# First get the minimum birth year for making a centered/scaled/whatever version
+# of the birth year variable.
+MIN_BY <- raw_data$dateofbirth |> substring(1, 4) |> as.integer() |> min()
+
+# Now we make those changes in the dataset.
+dat_model <-
+	dat_clean |>
+	dplyr::mutate(
+		# We need to make versions of the time and birth year variables that are
+		# closer to being scale-free than the current versions -- models that
+		# have those numbers in the thousands have worse conditioning problems that
+		# can lead to numerical issues in an otherwise fine model. So we'll scale
+		# the year variable by subtracting 2013, the first year, and we'll scale
+		# the birth_year variable by subtracting the minimum birth year.
+		year_c = year - 2013,
+		birth_year_c = birth_year - MIN_BY,
+		# We'll also center the age variable by subtracting 65.
+		age_c = age - 65
+	) |>
+	# Finally let's reorder the variables. This is for no practical purpose
+	# but it makes me happier.
+	dplyr::select(
+		id, study, season, year, year_c, age, age_c, birth_year, birth_year_c,
+		dose, vaccine_name, strain_name, strain_type, pretiter, log_pretiter,
+		posttiter, log_posttiter, fold_change, titer_increase, seroconversion,
+		seroprotection
+	)
+
+# Save data to file ============================================================
+readr::write_rds(
+	dat_model,
+	file = here::here("data", "processed", "model-data.Rds"),
+	compress = "gz"
+)
+readr::write_csv(
+	dat_model,
+	file = here::here("data", "processed", "model-data.csv")
+)
+
+# END OF FILE ==================================================================