Pooling visualization.Rmd

---
title: 'Pooling: Visualization'
author: "Xianbin Cheng"
date: "January 24, 2019"
output: html_document
---

# Method

1. Load all the necessary packages and source codes.

```{r, message=FALSE, warning = FALSE}
library(tidyverse)
library(Hmisc)
source("1D_pooling_simulation.R")
source("2D_pooling_simulation.R")
source("STD_simulation.R")
```

2. Read rds files and necessary parameters.

  * These files contain simulation results of STD(48; 7; 4) pooling, one-dimensional and two-dimensional pooling. The iteration time is 1000. The seed is set at 123.

```{r}
pool_STD = readRDS(file = "STD_48_7_4_1000.rds")
pool_row = readRDS("Pool_row_48_1000.rds")
pool_col = readRDS("Pool_col_48_1000.rds")
pool_2d = readRDS("Pool_2d_48_1000.rds")

n = 48
q_val = 7
k_val = 4
thresh = 20

n_pos_vals = 1:47
n_iter = 1000
```

3. Clean them up for visualization.

```{r}
# STD
metrics_STD = do.call(what = rbind.data.frame, args = pool_STD) %>%
  gather(data = ., key = "Type", value = "Value", - n_pos)

cost_STD = do.call(what = rbind.data.frame, args = pool_STD) %>%
  mutate(putative_pos = sensi*n_pos + (1-speci)*(n-n_pos),
         n_test_total = q_val*k_val + putative_pos,
         n_pipette = q_val*k_val*q_val + k_val*q_val + putative_pos)
```

```{r}
# Pool by rows
metrics_row = do.call(what = rbind.data.frame, args = pool_row) %>%
  gather(data = ., key = "Type", value = "Value", - n_pos)
cost_row = do.call(what = rbind.data.frame, args = pool_row) %>%
  mutate(putative_pos = sensi*n_pos + (1-speci)*(n-n_pos),
         n_test_total = 6 + putative_pos,
         n_pipette = n + 6 + putative_pos)

# Pool by columns
metrics_col = do.call(what = rbind.data.frame, args = pool_col) %>%
  gather(data = ., key = "Type", value = "Value", - n_pos)
cost_col = do.call(what = rbind.data.frame, args = pool_col) %>%
  mutate(putative_pos = sensi*n_pos + (1-speci)*(n-n_pos),
         n_test_total = 8 + putative_pos,
         n_pipette = n + 8 + putative_pos)
```

```{r}
# 2D Pool
metrics_2d = do.call(what = rbind.data.frame, args = pool_2d) %>%
  gather(data = ., key = "Type", value = "Value", - n_pos)

cost_2d = do.call(what = rbind.data.frame, args = pool_2d) %>%
  mutate(putative_pos = sensi*n_pos + (1-speci)*(n-n_pos),
         n_test_total = 6 + 8 + putative_pos,
         n_pipette = n * 2 + 6 + 8 + putative_pos)
```

```{r}
### All metrics
metrics_all = bind_rows(metrics_STD, metrics_row, metrics_col, metrics_2d) %>%
  mutate(Pooling = c(rep("STD", times = nrow(metrics_STD)),
                     rep("Row", times = nrow(metrics_row)),
                     rep("Column", times = nrow(metrics_col)),
                     rep("2D", times = nrow(metrics_2d))))

### All cost
cost_all = bind_rows(cost_STD, cost_row, cost_col, cost_2d) %>%
  mutate(Pooling = c(rep("STD", times = nrow(cost_STD)),
                     rep("Row", times = nrow(cost_row)),
                     rep("Column", times = nrow(cost_col)),
                     rep("2D", times = nrow(cost_2d))))
```

# Result

1. Visualization of metrics

```{r}
p1 = draw_metrics_all(df = metrics_all)
p1
```

2. Visualization of reagent cost.

```{r}
p2 = draw_cost_all(df = cost_all, n = n, var = n_test_total, ylab = "The total number of tests needed")
p2
```

3. Visualization of labor cost

```{r}
p3 = draw_cost_all(df = cost_all, n = n, var = n_pipette, ylab = "The number of transfers")
p3
```

```{r, eval = FALSE, echo= FALSE}
pdf(file = "Pooling Visualization.pdf")
  p1
  p2
  p3
dev.off()
```