Add figures for just TCGA data.

recount.Rmd

@@ -182,6 +182,7 @@ df_et_tcag_gtex$Type[df_et_tcag_gtex$Type == 'Metastatic'] <- 'Primary Tumor'
 df_et_tcag_gtex_er <- df_et_tcag_gtex[complete.cases(df_et_tcag_gtex[ , 3]),]
 df_et_tcag_gtex_er_type <- df_et_tcag_gtex_er[complete.cases(df_et_tcag_gtex_er[ , 4]),]
 df_et_tcag_gtex_er_type_com <- df_et_tcag_gtex_er_type[which(df_et_tcag_gtex_er_type$ER %in% c('Negative','Positive','Normal')),]
+df_et_tcag_gtex_type <- df_et_tcag_gtex[complete.cases(df_et_tcag_gtex[ , 4]),]
 
 # Seperate into PAM and type for different genes
 df_et_tcag_gtex_er_type_com$PARP1 <- log2(df_et_tcag_gtex_er_type_com$PARP1)
@@ -222,3 +223,43 @@ dev.off()
 
 ddx21_et.aov = aov(DDX21 ~ ER + Type, data=df_ddx21_et_tcag_gtex_er_type_summary)
 summary(ddx21_et.aov)
+
+### Type
+df_et_tcag_type <- df_et_tcag_gtex_type[which(df_et_tcag_gtex_type$Type %in% c('Primary Tumor','Solid Tissue Normal')),]
+df_et_tcag_type$PARP1 <- log(df_et_tcag_type$PARP1,2)
+df_et_tcag_type$DDX21 <- log(df_et_tcag_type$DDX21,2)
+
+## PARP1 and Type
+df_et_tcag_gtex_type_summary <- summarySE(df_et_tcag_type, measurevar="PARP1", groupvars=c("Type"))
+p <- ggplot(df_et_tcag_gtex_type_summary, aes(x = as.factor(Type), fill= Type)) +
+  geom_boxplot(aes(
+      lower = PARP1 - se,
+      upper = PARP1 + se,
+      middle = PARP1,
+      ymin = PARP1 - 2*se,
+      ymax = PARP1 + 2*se),
+    stat = "identity",lwd=1.5) + geom_hline(yintercept = mean(df_et_tcag_type$PARP1), linetype = 2) + labs(y="RPKM",x="PARP1") + theme_bw() + theme(panel.grid.major = element_blank(),panel.grid.minor = element_blank(),axis.ticks=element_line(size=1))
+jpeg('figures/PARP1_breast_correlation_nvc_boxplot.jpg')
+p
+dev.off()
+
+parp1_type.aov = aov(PARP1 ~ Type,data=df_et_tcag_type)
+summary(parp1_type.aov)
+
+## DDX21 and Type
+df_ddx21_et_tcag_gtex_type_summary <- summarySE(df_et_tcag_type, measurevar="DDX21", groupvars=c("Type"))
+
+p <- ggplot(df_ddx21_et_tcag_gtex_type_summary, aes(x = as.factor(Type), fill= Type)) +
+  geom_boxplot(aes(
+      lower = DDX21 - se,
+      upper = DDX21 + se,
+      middle = DDX21,
+      ymin = DDX21 - 2*se,
+      ymax = DDX21 + 2*se),
+    stat = "identity",lwd=1.5) + geom_hline(yintercept = mean(df_et_tcag_type$DDX21), linetype = 2) + labs(y="RPKM",x="DDX21") + theme_bw() + theme(panel.grid.major = element_blank(),panel.grid.minor = element_blank(),axis.ticks=element_line(size=1)) +  scale_y_continuous(limits=c(3.8,4.9))
+jpeg('figures/DDX21_breast_correlation_nvc_boxplot.jpg')
+p
+dev.off()
+
+ddx21_type.aov = aov(DDX21 ~ Type,data=df_et_tcag_type)
+summary(ddx21_type.aov)