tcga commit

notebooks/TCGA.qmd

@@ -319,12 +319,15 @@ ggsave(paste0(path,"number_mut_tp53_per_tumor_stage.jpeg"), device = "jpeg")
 
 # Linking expression and tumor stage
 
+We also can link the expression data (RNA-Seq) with the tumor stage data from the clinical metadata. [Check the TCGA RNA-Seq protocol online]{.aside}
+
 ```{r}
 rnaseq = readData[[2]]
 
 rnaseq
 
 ```
+As before, we need to shorten the sample IDs so they can match to the clinical data.
 
 ```{r}
 
@@ -345,6 +348,8 @@ rnaseq <- rnaseq[,!idx]
 
 ```
 
+Then, we can use the `limma` package to proceed with the differential expression analysis. In this case, we'll ise the tumor stage as a variable to explain the expression of the genes. 
+
 ```{r}
 
 library(limma)
@@ -356,6 +361,8 @@ topTable(ef1) |>
 
 ```
 
+Let's visualize two of the most expressed genes.
+
 ```{r}
 par(mfrow=c(1,2))
 boxplot(split(assay(rnaseq)["PAM",], rnaseq$t_stage), main="PAM")    # higher expression in lower t_stage
@@ -364,6 +371,10 @@ boxplot(split(assay(rnaseq)["PAIP2",], rnaseq$t_stage), main="PAIP2")
 
 # Linking methylation and expression
 
+Finally, we can use the methylation data with the expression data. This is important because methylated cytosines of the DNA change the expression of the genes.
+
+Some patients have methylation data for multiples tissue types. This information is encoded in the fourth component of the sample names. The code `01A` correspond to primary tumor samples and the code `11A` correspond to normal tissue. We'll keep the primary tumor samples.
+
 ```{r}
 
 methyl <- readData[[3]]
@@ -381,11 +392,15 @@ methyl = methyl[,idx]
 methyl
 ```
 
+As before, let's truncate the names of the sample to match the clinical data.
+
 ```{r}
 colnames(methyl) <- colnames(methyl) |> 
   str_sub(start = 1,end = 12)
 ```
 
+We can add the clinical data to the methyl object and count the number of patients with methylation and transcription data.
+
 ```{r}
 colData(methyl) <- as(clin[colnames(methyl),],"DataFrame")
 
@@ -394,6 +409,8 @@ intersect(colnames(methyl), colnames(rnaseq)) |> length()
 
 ```
 
+Let's subset common sample names and check the methylation data as row ranges.
+
 ```{r}
 
 methyl_subset = methyl[,which(colnames(methyl) %in% colnames(rnaseq))]
@@ -404,6 +421,8 @@ methyl_genes
 
 ```
 
+This function takes a gene symbol and returns a scatter plot showing the relationship between 3 different sites near that gene and gene expression.
+
 ```{r}
 
 #| label: figplot
@@ -444,4 +463,3 @@ me_rna_cor("TAC1", mpick=3)
 
 ```
 
-## Conclusion