Estimations_Inferences_2.Rmd

---
title: "problem_set_5"
author: "Lydia Holley"
date: "2024-02-12"
output: pdf_document
---

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

library(dplyr)
library(ggplot2)
library(AICcmodavg)


# import data
gene_info<-read.csv("GLDS-207_rna_seq_ERCCnorm_differential_expression.csv", header = FALSE)
sample_data<-read.csv("OSD-207-samples.csv")

gene_info = setNames(data.frame(t(gene_info[,-1])), gene_info[,1])

```

# Step 1

```{r join datasets}

joined<-inner_join(sample_data, gene_info, join_by(Sample_Name == ENSEMBL))


```

## Question 1 - How many flies have the genotype "CS" and gene expression for the gene "FBgn0000015" greater than 1

-   5 flies have genotype "CS" and gene expression for the gene "FBgn0000015" greater than 1

```{r q1}

gnCS_g1<-joined[joined$Factor_Value._Genotype %in% c("CS"), ]
gnCS_g1<-gnCS_g1[gnCS_g1$FBgn0000015 > 1, ]
nrow(gnCS_g1)

```

# Step 2

## *Step 2.1 Multiple Tests Correction*

```{r mtc}

alpha<-(0.05/13885)

```

## Question 2 - Report the adjusted p-value for this test

-   the adjusted p-value is 3.601e-06

## *Step 2.2 - Test all genes using ANOVA*

```{r multiple anova}
alpha<-(0.05/13885)

colSeq<-seq(from = 16, to = 13900, by = 1)
q3_list<-list()

for (i in colSeq) {
  column<-joined[[i]]
  column_name<-colnames(joined[i])
  anova<-aov(as.numeric(column) ~ joined$Factor_Value._Spaceflight)
  p_value<-summary(anova)[[1]][["Pr(>F)"]][1]
  if (p_value<alpha){
    q3_list<-c(q3_list,column_name)
  }
}

length(q3_list)

```

## Question 3 - Report the total number of genes with a statistically significant difference in mean gene expression

-   464 genes have a statistically significant difference in mean gene expression

# Step 3 Investigate the gene FBgn0038749

## *Step 3.1 Visualize our data*

## Question 4 - Create the boxplot

```{r boxplot}

joined$FBgn0038749<-as.numeric(joined$FBgn0038749)
anova<-aov(joined$FBgn0038749 ~ joined$Factor_Value._Spaceflight)
summary(anova)

ggplot(joined, aes(x=Factor_Value._Spaceflight, y=FBgn0038749, color=Factor_Value._Spaceflight)) + stat_boxplot(width=0.1)+geom_jitter( aes(fill=Factor_Value._Spaceflight), size=5, alpha=0.5)+theme_bw()

```

## *Step 3.2 Test Variances*

```{r bartlett test}

bartlett.test(FBgn0038749 ~ Factor_Value._Spaceflight, data=joined)

```

## Question 5 - Does our dataset violate the assumption of homogeneity?

-   our dataset does not violate the assumption of homogeneity

## *Step 3.4 Test residuals for normality*

```{r normality}
anova<-aov(joined$FBgn0038749 ~ joined$Factor_Value._Spaceflight)
residuals<-anova$residuals
shapiro.test(residuals)

```

## Question 6 - Does our dataset violate the assumption of normality?

-   our data does violate the assumption of normality

## *Step 3.5 Test for the impact of genotype*

```{r genotype}

# Two two-way anovas w/ spaceflight/control and genotype
# one do not predict interaction between the variables
two_way_add<-aov(FBgn0038749 ~ Factor_Value._Spaceflight + Factor_Value._Genotype, data = joined)
print(summary(two_way_add))

# second allow for interaction between the variables
two_way_inter<-aov(FBgn0038749 ~ Factor_Value._Spaceflight * Factor_Value._Genotype, data = joined)
print(summary(two_way_inter))

# Generate and use the AIC table to report the best model.
model.set<-list(two_way_add, two_way_inter)
model.names<-c("additive","interaction")
aictab(model.set, modnames=model.names)

```

## Question 7 - Is there an interaction between space-flight/ground-control and genotype?

-   the data suggests there is not an interaction between space-flight/ground-control and genotype

## *Step 3.6 Find out the gene name*

```{r gene name}
gene<-gene_info$FBgn0038749
gene[2]

```

## Question 8 - What is the gene name of FBgn0038749

-   "exit protein of rhodopsin and TRP A"

# *Step 4 Investigate the gene FBgn0033687*

## Question 9 - Create the boxplot

```{r boxplot2}

joined$FBgn0033687<-as.numeric(joined$FBgn0033687)
anova<-aov(joined$FBgn0033687 ~ joined$Factor_Value._Spaceflight)
summary(anova)

ggplot(joined, aes(x=Factor_Value._Spaceflight, y=FBgn0033687, color=Factor_Value._Spaceflight)) + stat_boxplot(width=0.1)+geom_jitter( aes(fill=Factor_Value._Spaceflight), size=5, alpha=0.5)+theme_bw()

```

## *Step 3.2 Test Variances*

```{r variances}

bartlett.test(FBgn0033687 ~ Factor_Value._Spaceflight, data=joined)

```

## Question 10 - Does our dataset violate the assumption of homogeneity?

-   our data does not violate the assumption of homogeneity

## *Step 3.4 Test residuals for normality*

```{r residuals test}

anova<-aov(joined$FBgn0033687 ~ joined$Factor_Value._Spaceflight)
residuals<-anova$residuals
shapiro.test(residuals)

```

## Question 11 - Does our dataset violate the assumption of normality?

-   our data does violate the assumption of normality

## *Step 3.5 Test for the impact of genotype*

```{r impact}
# Two two-way anovas w/ spaceflight/control and genotype
# one do not predict interaction between the variables
two_way_add<-aov(FBgn0033687 ~ Factor_Value._Spaceflight + Factor_Value._Genotype, data = joined)
print(summary(two_way_add))

# second allow for interaction between the variables
two_way_inter<-aov(FBgn0033687 ~ Factor_Value._Spaceflight * Factor_Value._Genotype, data = joined)
print(summary(two_way_inter))

# Generate and use the AIC table to report the best model.
model.set<-list(two_way_add, two_way_inter)
model.names<-c("additive","interaction")
aictab(model.set, modnames=model.names)

```

## Question 12 - Is there an interaction between space-flight/ground-control and genotype?

-   the data suggests there is an interaction between space-flight/ground-control and genotype

## *Step 3.6 Find out the gene name*

```{r name}
gene<-gene_info$FBgn0033687
gene[2]

```

## Question 13 - What is the gene name of FBgn0033687

-   "uncharacterized protein"