---
title: "SDY144: Correlation of HAI/Virus Neutralizition Titer and Cell Counts"
author: "Renan Sauteraud"
date: "`r Sys.Date()`"
output:
  html_document:
    toc: true
    toc_float: true
    number_sections: true
vignette: >
  %\VignetteIndexEntry{SDY144: Correlation of HAI/Virus Neutralizition Titer and Cell Counts}
  %\VignetteEngine{knitr::rmarkdown}
--- 

```{r knitr, echo = FALSE}
library(knitr)
opts_chunk$set(message = FALSE, fig.align = "center", fig.width = 10, fig.height = 8)
```

```{r netrc_req, echo = FALSE}
# This chunk is only useful for BioConductor checks and shouldn't affect any other setup
if (!any(file.exists("~/.netrc", "~/_netrc"))) {
    labkey.netrc.file <- ImmuneSpaceR:::get_env_netrc()
    labkey.url.base <- ImmuneSpaceR:::get_env_url()
}
```



***Correlations between hemagglutination inhibition (HI) and viral neutralization (VN) titers and plasmablast and plasma B cells among trivalent inactivated influenza vaccine (TIV) vaccinees.***

This reports reproduces Figure 2 of [Cao RG et al(2014)](http://www.jid.oxfordjournals.org/cgi/pmidlookup?view=long&pmid=24495909) published as part of the original study.



# Create a connection to SDY144

First, we initialize the connection to SDY144 using `CreateConnection`.

```{r}
library(ImmuneSpaceR)
library(data.table)
library(ggplot2)

con <- CreateConnection("SDY144")
```



# Retrieve and manipulate data

We grab the datasets of interests with the `getDataset` method.

```{r}
flow <- con$getDataset("fcs_analyzed_result", maxRows = 10000) # maxRows for Rlabkey 2.1.136 / LK 17.2
hai <- con$getDataset("hai")
vn <- con$getDataset("neut_ab_titer")
```

Then, we select the cell populations and time points of intereset.

```{r subset}
pb <- flow[population_name_reported %in% c("Plasma cells,Freq. of,B lym CD27+",
                                           "Plasmablast,Freq. of,Q3: CD19+, CD20-")]
pb <- pb[, population_cell_number := as.numeric(population_cell_number)]
pb <- pb[study_time_collected == 7 & study_time_collected_unit == "Days"] # 13 subjects
pb <- pb[, list(participant_id, population_cell_number, population_name_reported)]
```

We compute the HI and VN titer as the fold-increase between baseline and day 30.

```{r FC}
# HAI
hai <- hai[, response := value_preferred / value_preferred[study_time_collected == 0],
           by = "virus,cohort,participant_id"][study_time_collected == 30]
hai <- hai[, list(participant_id, virus, response)]
dat_hai <- merge(hai, pb, by = "participant_id", allow.cartesian = TRUE)

# VN
vn <- vn[, response:= value_preferred/value_preferred[study_time_collected == 0],
         by = "virus,cohort,participant_id"][study_time_collected == 30]
vn <- vn[, list(participant_id, virus, response)]
dat_vn <- merge(vn, pb, by = "participant_id", allow.cartesian = TRUE)
```



# Visualize using `ggplot2`

## Figure 2A

***Correlation between the absolute number of plasmablasts and plasma B cells 7 days after vaccination with and fold-increase of HI titers from baseline to day 30 after vaccination.***

```{r HAI, dev='png'}
ggplot(dat_hai, aes(x = population_cell_number, y = response)) +
  geom_point() + 
  geom_smooth(method = "lm") +
  facet_grid(virus ~ population_name_reported, scale = "free") +
  xlab("Number of cells") + 
  ylab("HI fold-increase Day 30 vs. baseline") + 
  theme_IS()
```


## Figure 2B

***Correlation between the absolute number of plasmablasts and plasma B cells 7 days after vaccination with and fold-increase of VN titers from baseline to day 30 after vaccination.***

```{r VN, dev='png'}
ggplot(dat_vn, aes(x = population_cell_number, y = response)) +
  geom_point() + 
  geom_smooth(method = "lm") +
  facet_grid(virus ~ population_name_reported, scale = "free") +
  xlab("Number of cells") + 
  ylab("VN fold-increase Day 30 vs. baseline") + 
  theme_IS()
```