Title

Computational Insights on Kidney Dysfunction and its Bodily Impacts

Background

Unilateral Ureteral Obstruction (UUO) is a commonly used experimental model for studying renal injury. While it effectively models the kidney's response to urinary tract blockage, its applications extend beyond this. UUO is an excellent model for investigating irreversible acute kidney disease and the mechanisms underlying chronic kidney diseases 1

What happens when the kidney is dysfunctional?

This project aims to answer this question by exploring genes significantly expressed in the presence of chronic diseases, and its overall impact on the body's processes.

This project is divided into these sections:

• Project Staging: This section details project dependencies, data loading and preparation, quality control, filtering, and normalization steps.
• Analysis: The analysis includes differential expression analysis and functional enrichment analysis.
• Results and Conclusion: The biological implications of the analyses were discussed.

Data Sources and Analyses

The datasets used in this project were obtained from an experimental study titled Silencing SMOC2 ameliorates kidney fibrosis by inhibiting fibroblast to myofibroblast transformation. The datasets are publicly available under the accession number GSE85209. The sequencing data for the cell lines can be accessed through the following accession numbers: GSM2260466 GSM2260467 GSM2260468 GSM2260469 GSM2260470 GSM2260471 GSM2260472

In this project, I will explore the genes differentially expressed in the Unilateral Ureteral Obstruction (UUO) mouse model.

The following analyses will be performed:

• Differential Expression Analysis (DEA): This analysis will focus on identifying the genes significantly expressed in the UUO mouse and control models. Four replicates of the UUO mice and three replicates of the healthy mouse models will be analyzed.

• Functional Enrichment Analysis (FEA): This analysis will determine the biological significance of the differentially expressed genes. By categorizing these genes into known biological functions and pathways, FEA will help uncover the molecular mechanisms and pathways involved in the disease state. This can provide insights into the processes involved and potentially lead to the identification of therapeutic targets for intervention.

  Project Staging

  This section is broken into three sub-sections:

• Project Dependencies: Import the necessary libraries for downstream analyses.

• Import and Prepare Data: Import raw datasets and prepare them for downstream processing.

• Quality Control, Filtering, and Normalization: Applying quality control measures and data transformations to generate high-quality, reliable data for downstream analysis.

Project dependencies

To start, we will need to import the R packages needed for the project

library(dplyr)
library(tidyverse)
library(pheatmap)
library(DESeq2)
library(ggplot2)
library(clusterProfiler)
library(org.Mm.eg.db)
library(pathview)
library(gage)
library(annotables)

Data import and preparation

#Import normal raw counts data

#Loading Raw counts from normal

n_raw_1 <- read.delim("C://Users//ghagi//Desktop//February//GSM2260466_SMOC2_normal.txt",  header = TRUE, sep = "\t", col.names = c("Gene_id", "normal_1"), stringsAsFactors = FALSE)
n_raw_2 <- read.delim("C://Users//ghagi//Desktop//February//GSM2260467_SMOC2_normal_3.txt", header = TRUE, sep = "\t", col.names = c("Gene_id", "normal_2"), stringsAsFactors = FALSE)
n_raw_3 <- read.delim("C://Users//ghagi//Desktop//February//GSM2260468_SMOC2_normal_4.txt", header = TRUE, sep = "\t", col.names = c("Gene_id", "normal_3"), stringsAsFactors = FALSE)

#Loading counts of uuo data
uuo_raw <- read.delim("C://Users//ghagi//Desktop//February//GSM2260469_SMOC2_UUO.txt", header = TRUE, sep = "\t", col.names = c("Gene_id", "uuo_1"),  stringsAsFactors = FALSE)
uuo_raw_2 <- read.delim("C://Users//ghagi//Desktop//February//GSM2260470_SMOC2_UUO_2.txt", header = TRUE, sep = "\t", col.names = c("Gene_id", "uuo_2"), stringsAsFactors = FALSE)
uuo_raw_3 <- read.delim("C://Users//ghagi//Desktop//February//GSM2260471_SMOC2_UUO_3.txt", header = TRUE, sep = "\t", col.names = c("Gene_id", "uuo_3"),  stringsAsFactors = FALSE)
uuo_raw_4 <- read.delim("C://Users//ghagi//Desktop//February//GSM2260472_SMOC2_UUO_4..txt", header = TRUE, sep = "\t", col.names = c("Gene_id", "uuo_4"),  stringsAsFactors = FALSE)

#Joining the tables for the normal counts
normal_raw <- n_raw_1 %>%
  left_join(n_raw_2, by= "Gene_id") %>%
  left_join(n_raw_3, by ="Gene_id")

rownames(normal_raw) <- normal_raw$Gene_id

#Join the tables for the uuo counts
uuo_raw <- uuo_raw %>%
  left_join(uuo_raw_2, by= "Gene_id") %>%
  left_join(uuo_raw_3, by ="Gene_id") %>%
  left_join(uuo_raw_4, by = "Gene_id")

#Joining the normal counts to the uuo count by the Gene Id
expression <- normal_raw %>%
  left_join(uuo_raw, by="Gene_id")

#Set rownames to Gene_id

rownames(expression) <- expression$Gene_id

#dropping the gene_id column
expression <- expression %>%
  select(-"Gene_id")


# Create metadata
meta_data <- data.frame(
  row.names = factor(c("normal_1", "normal_2", "normal_3", "uuo_1", "uuo_2", "uuo_3", "uuo_4")),
  genotype = factor(c("normal", "normal", "normal", "uuo", "uuo", "uuo", "uuo")),
  condition = factor(c("normal", "fibrosis", "normal", "fibrosis", "normal", "fibrosis", "normal")),
  replicate = c(1, 2, 3, 1, 2, 3, 4)
)


#View first 6 rows of raw counts

head(expression)

which outputs:

#create a deseq2 dataset

dds <- DESeqDataSetFromMatrix(
  countData = expression,
  colData = meta_data,
  design = ~genotype
)

#Normalizing the deseq counts
norm_dds <- estimateSizeFactors(dds)
sizeFactors(norm_dds)

dds_norm <- counts(norm_dds, normalized = TRUE)

#Variance stabilization tranformation
log_dds_norm <- vst(dds, blind = TRUE)

#extract the log tranformed counts
log_dds_norm_wt <- assay(log_dds_norm)

Quality control, Analysis and Filtering

#Plot a heatmap 
pheatmap(norm_cor, annotation_col= meta_data[, "genotype", drop=FALSE])

which outputs:

The healthy control group clustered together, while the uuo group clustered together showing a similarity in their gene expressions.

#PCA plot
plotPCA(log_dds_norm, intgroup = "genotype")

which outputs:

The PC1 and PC2 accounts for 95% of variance in the data, which is good for our analysis

#Create a Deseq object
deSeqd <- DESeq(dds)

#Dispersion plot
plotDispEsts(deSeqd)

which outputs:

#Extract results

norm_res <- results(deSeqd,
                    contrast  = c("genotype", "uuo", "normal") ,
                    alpha = 0.05, 
                    lfcThreshold = 0.5
)


# Join the results to the mouse reference genome and make it a dataframe
norm_res_all <- data.frame(norm_res) %>%
  rownames_to_column(var = "ensgene") %>%
  left_join(x = norm_res_all,
            y = grcm38[, c("ensgene", "symbol", "description")],
            by = "ensgene")

#Extract significantly expressed genes. My padj was set to 0.05 
significant_genes <- subset(norm_res_all, padj <0.05)

significant_genes <- significant_genes %>%
  arrange(padj)

# Filter DEGs based on adjusted p-value < 0.05 and log2 fold change > 0.5 or < -0.5
degs <- norm_res_all[norm_res_all$padj < 0.05 & (norm_res_all$log2FoldChange > 0.5 | norm_res_all$log2FoldChange < -0.5), ]

#Functional Enrichment Analysis (FEA)

#BP
go_results_BP <- enrichGO(gene         = degs$ensgene,
                       OrgDb        = org.Mm.eg.db,
                       keyType      = "ENSEMBL",
                       ont          = "BP", # Biological Process
                       pAdjustMethod = "BH",
                       pvalueCutoff  = 0.01,
                       qvalueCutoff  = 0.05,
                       readable     = TRUE)
#MF
go_results_MF <- enrichGO(gene         = degs$ensgene,
                          OrgDb        = org.Mm.eg.db,
                          keyType      = "ENSEMBL",
                          ont          = "MF", # Molecular Function
                          pAdjustMethod = "BH",
                          pvalueCutoff  = 0.01,
                          qvalueCutoff  = 0.05,
                          readable     = TRUE)
#CC
go_results_CC <- enrichGO(gene         = degs$ensgene,
                          OrgDb        = org.Mm.eg.db,
                          keyType      = "ENSEMBL",
                          ont          = "CC", # Cellular compartment
                          pAdjustMethod = "BH",
                          pvalueCutoff  = 0.01,
                          qvalueCutoff  = 0.05,
                          readable     = TRUE)

top_20_MF <- go_results_MF[1:20, ]


top_20_BP <- go_results_BP[1:20,]

top_20_CC <- go_results_CC[1:20, ]


ggplot(top_20_BP, aes(x = reorder(Description, -p.adjust), y = -log10(p.adjust))) +
  geom_bar(stat = "identity", fill = "skyblue") +
  coord_flip() +
  labs(title = "Top 20 BP",
       x = "Biological Processes",
       y = "-log10(p.adjust)", 
       caption  = "MikeHail, 2025") +
  theme_minimal()

which outputs:

ggplot(top_20_CC, aes(x = reorder(Description, -p.adjust), y = -log10(p.adjust))) +
  geom_bar(stat = "identity", fill = "skyblue") +
  coord_flip() +
  labs(title = "Top 20 CC",
       x = "Cellular Compartment",
       y = "-log10(p.adjust)",
       caption = "MikeHail, 2025") +
  theme_minimal()

which outputs:

ggplot(top_20_MF, aes(x = reorder(Description, -p.adjust), y = -log10(p.adjust))) +
  geom_bar(stat = "identity", fill = "skyblue") +
  coord_flip() +
  labs(title = "Top 20 MF",
       x = "Molecular Function",
       y = "-log10(p.adjust)", 
       caption = "MikeHail, 2025") +
  theme_minimal()

which outputs:

#Pathway Analysis

# Set up the biomart
ensembl <- useMart("ensembl", dataset = "mmusculus_gene_ensembl")

# Convert gene symbols to Entrez IDs
gene_conversion <- getBM(attributes = c("ensembl_gene_id", "entrezgene_id"),
                         filters = "ensembl_gene_id",
                         values = significant_genes$ensgene,
                         mart = ensembl)

# Use converted Entrez IDs for KEGG analysis
entrez_ids <- gene_conversion$entrezgene_id



# Perform KEGG pathway enrichment analysis
kegg_results <- enrichKEGG(gene = entrez_ids,
                           organism = "mmu", # "mmu" stands for Mus musculus (mouse)
                           keyType = "kegg",
                           pvalueCutoff = 0.05)

# kegg results into dataframe

kegg_results <- data.frame(kegg_results) 
top_20_kegg <- kegg_results[1:20,]
  


# Visualize pathway results

ggplot(top_20_kegg, aes(x= reorder(Description, -p.adjust), y= -log10(p.adjust)))+
  geom_bar(stat = "identity", fill= "skyblue") +
  coord_flip() +
  labs(x = "Description", y= "-10logpadjust",
       caption = "MikeHail, 2025")+
  theme_classic()

which outputs:

Results and Conclusion

Functional Enrichment Analysis

• Biological processes

The primary biological processes identified are centered around energy production and the metabolism of molecules, ultimately leading to energy generation. These processes highlight changes in how energy is utilized within the cell. Additionally, terms related to leukocyte migration, cell-to-cell adhesion, and immune response-regulating signaling pathways indicate potential alterations in immune responses.

• Molecular function

Furthermore, the enriched molecular functions involve the movement of molecules across membranes, binding interactions between molecules, catalytic activities involving oxidation-reduction reactions, and roles in immune responses.

• Cellular compartment

The enriched cellular compartments (CC) include the mitochondrial respiratory chain complexes, which are directly involved in ATP production via oxidative phosphorylation, and the mitochondrial matrix, which hosts the Krebs cycle and fatty acid oxidation. The NADH dehydrogenase complex is a critical entry point for electrons into the respiratory chain, essential for ATP synthesis. Other key compartments are the oxidoreductase complexes, plasma membrane regions (involved in absorption/secretion and signaling/trafficking processes), peroxisomes (responsible for lipid breakdown and detoxification, e.g., β-oxidation, especially in liver/kidney cells), and the extracellular matrix (primarily maintaining structural integrity).

Hence, the enriched compartments highlight the critical roles in energy production, trafficking, signaling, secretion/absorption, and lipid breakdown and detoxification.

Pathway Enrichment Analysis

The enriched pathway analysis revealed significant associations with:

• Energy dysregulation (e.g., oxidative phosphorylation, thermogenesis)

• Metabolic and organ-specific disorders (e.g., diabetic cardiomyopathy, non-alcoholic fatty liver disease)

• Neurodegenerative diseases (e.g., Alzheimer's disease, Parkinson's disease, prion disease)

• Oxidative stress (e.g., ROS-linked carcinogenesis)

• Immune/inflammatory signaling (e.g., cytokine interactions, complement cascades)

• Structural/cellular maintenance (e.g., ECM-receptor interaction, cytoskeleton remodeling)

Conclusion

The overall analysis revealed that the differentially expressed genes in UUO mice models are associated with metabolic disorders, neurodegenerative diseases, inflammatory signaling, and alterations in cellular integrity. There is a remodeling of the kidney structure, which suggests conditions that include fibrosis. Fibrosis is the formation of excess fibrous connective tissue, which impairs the normal kidney architecture. Chronic kidney diseases might exacerbate or contribute to these conditions, highlighting the overall negative impacts that kidney disease has on health.