DSiB_project2.Rmd

---
title: "Data Science for Bionformatics - project 2"
output:
  html_document:
    theme: readable
editor_options:
  chunk_output_type: console
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```
# Load libraries

```{r, message = FALSE}
library('ggplot2') 
library('ggthemes') 
library('scales')
library('grid') 
library('gridExtra')
library('corrplot')
library('ggraph') 
library('igraph') 
library('dplyr') 
library('nycflights13')
library('readr') 
library('tibble') 
library('tidyr')
```

# Import data

#My assignment is dealing with data on cancer incidence. Since a cancer tumor might have thousands of genetic mutations, it's critical to distinguish between those that help the tumor grow and those that don't.Now I analyse and compare the mutation types of 2 datasets based on experiments and on assumption.

```{r, message=FALSE, warning=FALSE}

train <- read_csv('C:/Users/Tulajdonos/Documents/DSiB/training_variants.csv')
test  <- read_csv('C:/Users/Tulajdonos/Documents/DSiB/test_variants.csv')

```
# Overview the data
```{r}
train <- train %>%
  mutate(Gene = factor(Gene),
         Variation = factor(Variation),
         Class = factor(Class))

test <- test %>%
  mutate(Gene = factor(Gene),
         Variation = factor(Variation))

summary(train)

print(train)
print(test)
```
#calculating how many different Genes are in the train dataset -> 264 rows -> 264 different genes
```{r}
train %>%
  group_by(Gene) %>%
  summarise(th = n()) %>%
  arrange(desc(th))
```

#calculating how many different Genes are in the test dataset -> 1397 rows -> 1397 different genes

```{r}
test %>%
  group_by(Gene) %>%
  summarise(th = n()) %>%
  arrange(desc(th))
```

#we can see that the test dataset contains more data which is because they wanted to avoid hand labeling, so the test dataset contains auto-generated data too

#now I want to count the different variations for each dataset

```{r}
train %>%
  group_by(Variation) %>%
  summarise(th = n()) %>%
  arrange(desc(th))
```
#number of variations in the train dataset = 2996

```{r}
test %>%
  group_by(Variation) %>%
  summarise(th = n()) %>%
  arrange(desc(th))
```
#number of variations in the train dataset = 5628

# Visualization

#checking the distribution of Class from the train dataset

```{r}
train %>%
  ggplot(aes(Class)) +
  geom_bar(fill = "#FF6600")
```

#the class 3,8,9 have the lowest frequency value and apparently class 7 stands out, so it's the most common one while the remaining classes can be classified as medium frequent classes


#I want to visualize how the most common genes are distributed:

```{r}
freq_gene <- train %>%
  group_by(Gene) %>%
  summarise(th = n()) %>%
  filter(th > 25)

freq_gene %>%
  ggplot(aes(reorder(Gene, -th), th)) +
  geom_point(size = 3,color='darkblue') +
  ggtitle("Frequency distribution \n of the most common genes in the train dataset") +
  labs(x = "Gene", y = "Frequency") 
```

```{r}
freq_gene_test <- test %>%
  group_by(Gene) %>%
  summarise(th = n()) %>%
  filter(th > 25)

freq_gene_test %>%
  ggplot(aes(reorder(Gene, -th), th)) +
  geom_point(size = 3,colour='red') +
  ggtitle("Frequency distribution \n of the most common genes in the test dataset") +
  labs(x = "Gene", y = "Frequency")
```

#From the graphs, we can deduce that the test data contains fewer frequent genes.

#furthermore I wanted to compare the 2 datasets based on variations (types of the mutations) -> I removed the ID column from the test dataset and ID and Class from the train dataset

```{r}
train_gene_var <- train %>% mutate(dataset = factor("train")) %>% select(-Class, -ID)
test_gene_var <- test %>% mutate(dataset = factor("test")) %>% select(-ID)

compare_g_v <- full_join(train_gene_var, test_gene_var)

compare_g_v %>%
  group_by(Variation, dataset) %>%
  summarise(th = n()) %>%
  filter(th > 3) %>%
  ggplot(aes(reorder(Variation, -th), th, colour = dataset)) +
  geom_point(size = 3) +
  ggtitle("Comparison of the most frequent mutation types in the 2 datasets") +
  coord_cartesian(ylim = c(0, 100)) +
  labs(x = "Variation", y = "Frequency")
```