---

title: "Run GB on Combined T cells - surface"

author: Tobias Roider

date: "Last compiled on `r format(Sys.time(), '%d %B, %Y, %X')`"

output: 

  rmdformats::readthedown: 

editor_options: 

  chunk_output_type: console

---

```{r options, include=FALSE, warning = FALSE}

library(knitr)

opts_chunk$set(echo=TRUE, tidy=FALSE, include=TRUE, message=FALSE,

               dpi = 100, cache = FALSE, warning = FALSE)

opts_knit$set(root.dir = "../")

options(bitmapType = "cairo")

```

# Load packages

```{r Load packages and functions}

library(Seurat)

library(tidyverse)

library(caret)

library(rsample)

source("R/Functions.R")

```

# Read data

```{r read data, fig.width=4, fig.height=4}

sobj <- readRDS("output/Tcells_Integrated.rds")

sobj$IdentI <- factor(sobj$IdentI, levels = sobj$IdentI %>% unique %>% sort())

Idents(sobj) <- "IdentI"

DimPlot(sobj, label = T, reduction = "wnn.umap", raster = TRUE, 

        raster.dpi = c(500,500), group.by = "IdentI")+

  NoLegend()

```

# Split data using rsample package

```{r split data}

set.seed(1)

split <- sobj@assays$integratedADT@scale.data %>% t %>% data.frame() %>% 

  cbind(., FetchData(sobj, vars = c("IdentI"))) %>% 

  select(-IdentI) %>% 

  initial_split(prop = 0.35)

train <- training(split) %>% as.matrix()

test <- testing(split) %>% as.matrix()

classes_split <- initial_split(as_tibble(Idents(sobj)), prop = 0.35)

classes_split[["in_id"]] <- split[["in_id"]] # Ensures corresponding celltypes

classes_training <- training(classes_split) %>% .$value %>% as.factor()

classes_test  <- testing(classes_split) %>% .$value %>% as.factor()

# Ensure all classes are represented

all(levels(classes_training) == levels(classes_test))

```

# Train model

```{r train model, message=FALSE, warning=FALSE, results='hide'}

trControl <- trainControl(method = "cv", number = 10, allowParallel = TRUE,

                          search = "random", verboseIter = FALSE)

set.seed(23)

gb_fit <- train(x = train,

                y = classes_training,

                method = "xgbTree",

                importance = TRUE,

                metric = "Accuracy",

                maximize = TRUE,

                trControl = trControl)

```

# Print model

```{r print model}

print(gb_fit)

```

# Predict

```{r predict model}

predict_test <- predict(gb_fit, newdata = test)

confusionMatrix(predict_test, classes_test)

```

# Store classes

```{r store classes}

names(classes_training) <- rownames(train)

names(classes_test) <- rownames(test)

names(predict_test) <- rownames(test)

classes <- list(classes_training, classes_test, predict_test)

names(classes) <- c("train", "test", "predict")

```

# Show most important markers

```{r important markers}

varImp(gb_fit, scale = TRUE, top=40)

```

# Save objects

```{r save}

saveRDS(gb_fit, file = "output/GBresults_Combined_T_surface.rds")

saveRDS(classes, file = "output/GBclasses_Combined_T_surface.rds")

```

# Session info

```{r session info}

sessionInfo()

```