uses revised facet_grid with side panels

cohort1/.Rhistory

@@ -1,43 +1,182 @@
 plot()
+library(hexSticker)
+install.packages("hexSticker")
+library(hexSticker)
+sticker("./cellScaleFactorsIMAGE.png",
+package="cellScaleFactors",
+p_size=20, s_x=1, s_y=.75, s_width=.6,
+filename="cellScaleFactors.png")
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=20, s_x=1, s_y=.75, s_width=.6,
+filename="cellScaleFactors.png")
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=20, s_x=1, s_y=.75, s_width=.6,
+filename="cellScaleFactors.png")
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=15, s_x=1, s_y=.75, s_width=.6,
+filename="cellScaleFactors.png")
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=15, s_x=1, s_y=.75, s_width=.5,
+filename="cellScaleFactors.png")
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=15, s_x=1, s_y=.8, s_width=.5,
+filename="cellScaleFactors.png")
+?sticker
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=15, s_x=1, s_y=.8, s_width=.5,
+filename="cellScaleFactors.png",)
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=15, s_x=1, s_y=.8, s_width=.5,
+filename="cellScaleFactors.png", h_fill="#111111")
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=15, s_x=1, s_y=.8, s_width=.5,
+filename="cellScaleFactors.png", h_fill="#000000")
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=15, s_x=1, s_y=.8, s_width=.5,
+filename="cellScaleFactors.png", h_fill="#999999")
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=15, s_x=1, s_y=.8, s_width=.5,
+filename="cellScaleFactors.png", h_fill="#FCFCFC")
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=15, s_x=1, s_y=.8, s_width=.5,
+filename="cellScaleFactors.png", h_fill="#FCFCFC",
+p_color = "#111111")
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=15, s_x=1, s_y=.8, s_width=.5,
+filename="cellScaleFactors.png",
+h_fill="#FCFCFC", p_color = "#111111", h_color = "#FF8E60")
+sticker("./cellScaleFactorsIMAGE.PNG",
+package="cellScaleFactors",
+p_size=15, s_x=1, s_y=.8, s_width=.5,
+filename="cellScaleFactors.png",
+h_fill="#FCFCFC", p_color = "#111111", h_color = "#FF674F")
 setwd("C:/Users/User/Documents/GitHub/deconvo_method-paper/cohort1")
+knitr::opts_chunk$set(fig.width=10, fig.height=10, echo = F)
 libv <- c("ggplot2")
-sapply(libv, library, character.only = TRUE)
-knitr::opts_chunk$set(echo = TRUE)
+sapply(libv, library, character.only = T)
+# load env data
 setwd("..")
 setwd("..")
-list.files()
-load("./env/02_pseudobulk/02_k3.RData")
-new.plot
-# plot proportions multipanel -- scale vs with scale
-new.plot <- ggplot(dfp.ct, aes(x = true.noscale, y = pred.noscale)) +
-geom_point(size = 4, alpha = 0.5) +
-geom_abline(slope = 1, intercept = 0) +
-geom_hline(yintercept = 0.5) + geom_vline(xintercept = 0.5) + theme_bw() +
-xlab("Known") + ylab("Predicted") +
-xlim(0, 1) + ylim(0, 1) + facet_wrap(~celltype) +
+load("./env/05_bulk/01_run_manual_script.RData")
+# format plot data
+dfp <- df.k2
+# format variable states
+dfp$experiment.type <- paste0("z_type : ", dfp$experiment.type)
+dfp$assay.name.lutearg <- paste0("rescale_type : ", dfp$assay.name.lutearg)
+dfp$s.set.label <- paste0("s_set : ", dfp$s.set.label)
+dfp$bulk.sample.id <- gsub("c1", "c", dfp$bulk.sample.id)
+dfp$bulk.sample.id <- gsub("k1", "k", dfp$bulk.sample.id)
+dfp$bulk.sample.id <- gsub("o1", "o", dfp$bulk.sample.id)
+# experiment conditions
+dfp$bulk.sample.condition <- cd[dfp$bulk.sample.id,]$expt_condition
+table(dfp$bulk.sample.condition)
+# filter na
+dfp <- na.omit(dfp)
+expt <- data.frame(z.reference.type = c("shared", "within", "shared", "within"),
+y.expression.scale = c("counts", "counts", "RPKM", "RPKM"))
+expt <- do.call(rbind, lapply(seq(length(list.s.pred)), function(ii){
+expt.iter <- expt
+expt.iter
+}))
+expt$cell.label.type <- "k2"
+# table visualization
+knitr::kable(expt)
+data.dict <- list(z.reference.type = "Type of cell type reference for deconvolution",
+y.expression.scale = "Scale of the bulk sample data passed to lute",
+s.pred.set.name = "Label of the S factor set",
+s.pred.set.values = "S factor set values.",
+cell.label.type = "Cell type identifier")
+data.dict <- do.call(rbind, lapply(seq(length(data.dict)), function(ii){c(names(data.dict)[ii], data.dict[[ii]])}))
+knitr::kable(data.dict)
+# new plot
+new.plot.scatter <-
+ggplot(dfp, aes(x = true.neuron, y = neuron, color = bulk.scale.type)) +
+theme_bw() + geom_point(alpha = 0.5) + geom_abline(slope = 1, intercept = 0) +
+facet_wrap(experiment.type*assay.name.lutearg~s.set.label) +
+xlab("Known") + ylab("Predicted") + xlim(0, 1) + ylim(0, 1) +
 theme(axis.text.x = element_text(angle = 45, hjust = 1))
-new.plot
-# plot proportions multipanel -- scale vs with scale
-dfp.ct$celltype <- factor(dfp.ct$celltype, levels = c("Excit", "Inhib", "glial"))
-new.plot <- ggplot(dfp.ct, aes(x = true.noscale, y = pred.noscale)) +
-geom_point(size = 4, alpha = 0.5) +
-geom_abline(slope = 1, intercept = 0) +
-geom_hline(yintercept = 0.5) + geom_vline(xintercept = 0.5) + theme_bw() +
-xlab("Known") + ylab("Predicted") +
-xlim(0, 1) + ylim(0, 1) + facet_wrap(~celltype) +
+new.plot.scatter
+# new plot
+new.plot.scatter <-
+ggplot(dfp, aes(x = true.neuron, y = neuron, color = bulk.scale.type)) +
+theme_bw() + geom_point(alpha = 0.5) + geom_abline(slope = 1, intercept = 0) +
+facet_wrap(experiment.type+assay.name.lutearg~s.set.label) +
+xlab("Known") + ylab("Predicted") + xlim(0, 1) + ylim(0, 1) +
+theme(axis.text.x = element_text(angle = 45, hjust = 1))
+new.plot.scatter
+# new plot
+new.plot.scatter <-
+ggplot(dfp, aes(x = true.neuron, y = neuron, color = bulk.scale.type)) +
+theme_bw() + geom_point(alpha = 0.5) + geom_abline(slope = 1, intercept = 0) +
+facet_grid(experiment.type+assay.name.lutearg~s.set.label) +
+xlab("Known") + ylab("Predicted") + xlim(0, 1) + ylim(0, 1) +
+theme(axis.text.x = element_text(angle = 45, hjust = 1))
+new.plot.scatter
+# new plot
+new.plot.scatter <-
+ggplot(dfp, aes(x = true.neuron, y = neuron, color = bulk.scale.type)) +
+theme_bw() + geom_point(alpha = 0.5) + geom_abline(slope = 1, intercept = 0) +
+facet_grid(experiment.type~assay.name.lutearg+s.set.label) +
+xlab("Known") + ylab("Predicted") + xlim(0, 1) + ylim(0, 1) +
 theme(axis.text.x = element_text(angle = 45, hjust = 1))
-new.plot
+new.plot.scatter
 setwd("..")
 setwd("..")
-new.plot.name <- "fig2c_cohort1.jpg"
-jpeg(file.path("./figures/02_pseudobulk/", new.plot.name),
-width = 6, height = 3, units = "in", res = 400)
-new.plot
+new.plot.path <- "./figures/05_bulk/figs3a_scatter.jpg"
+jpeg(new.plot.path, width = 7.5, height = 7, res = 400, units = "in")
+new.plot.scatter
 dev.off()
 setwd("..")
 setwd("..")
-new.plot.name <- "fig2c_cohort1.jpg"
-jpeg(file.path("./figures/02_pseudobulk/", new.plot.name),
-width = 6.5, height = 3, units = "in", res = 400)
-new.plot
+load("./env/05_bulk/02_crossvalidate_script.RData")
+knitr::opts_chunk$set(fig.width=10, fig.height=10, echo = F)
+libv <- c("ggplot2")
+sapply(libv, library, character.only = T)
+setwd("..")
+setwd("..")
+load("./env/05_bulk/02_crossvalidate_script.RData")
+dfp <- df.k2
+# format plot data
+dfp <- dfp[dfp$crossvalidation=="validation",]
+dfp[dfp$s.set.label=="s.null",]$s.set.label <- "unscaled"
+dfp[dfp$s.set.label=="s.manual",]$s.set.label <- "scaled"
+dfp$experiment.type <- paste0("z_type : ", dfp$experiment.type)
+dfp$bulk.sample.id <- gsub("c1", "c", dfp$bulk.sample.id)
+dfp$bulk.sample.id <- gsub("k1", "k", dfp$bulk.sample.id)
+dfp$bulk.sample.id <- gsub("o1", "o", dfp$bulk.sample.id)
+# experiment conditions
+dfp$bulk.sample.condition <- cd[dfp$bulk.sample.id,]$expt_condition
+table(dfp$bulk.sample.condition)
+dfp <- na.omit(dfp)
+new.plot.scatter <- ggplot(dfp, aes(x = true.neuron, y = neuron)) +
+theme_bw() + geom_point(alpha = 0.5) +
+geom_abline(slope = 1, intercept = 0) +
+facet_wrap(~s.set.label*experiment.type) + xlab("Known") +
+ylab("Predicted") + xlim(0, 1) + ylim(0, 1) +
+theme(axis.text.x = element_text(angle = 45, hjust = 1))
+new.plot.scatter
+new.plot.scatter <- ggplot(dfp, aes(x = true.neuron, y = neuron)) +
+theme_bw() + geom_point(alpha = 0.5) +
+geom_abline(slope = 1, intercept = 0) +
+facet_grid(s.set.label~experiment.type) + xlab("Known") +
+ylab("Predicted") + xlim(0, 1) + ylim(0, 1) +
+theme(axis.text.x = element_text(angle = 45, hjust = 1))
+new.plot.scatter
+setwd("..")
+setwd("..")
+new.plot.path <- "./figures/05_bulk/figs2b_validate.jpg"
+jpeg(new.plot.path, width = 4.5, height = 5, units = "in", res = 400)
+new.plot.scatter
 dev.off()

cohort1/notebooks/02_summarize_mae/01_data_summaries.Rmd

@@ -100,6 +100,10 @@ dev.off()
 ```
 # Sample quality summaries
 
+```{r}
+sample.id.train
+```
+
 Bulk
 
 ```{r}
@@ -110,12 +114,35 @@ assays(bulk)[["counts"]] |> colSums() |> summary()
 
 # count zeros
 apply(assays(bulk)[["counts"]],2,function(ci){length(ci[ci==0])}) |> summary()
+
+length(intersect(bulk$batch.id2, sample.id.train)) 
+length(intersect(bulk$batch.id2, sample.id.validate))
+
+bf <- bulk[bulk$library_prep=="Bulk",]
+length(intersect(bf$batch.id2, sample.id.train)) 
+length(intersect(bf$batch.id2, sample.id.validate))
+
+bf <- bulk[bulk$library_prep=="Cyto",]
+length(intersect(bf$batch.id2, sample.id.train)) 
+length(intersect(bf$batch.id2, sample.id.validate))
+
+bf <- bulk[bulk$library_prep=="Nuc",]
+length(intersect(bf$batch.id2, sample.id.train)) 
+length(intersect(bf$batch.id2, sample.id.validate))
+
+bf <- bulk[bulk$library_type=="polyA",]
+length(intersect(bf$batch.id2, sample.id.train)) 
+length(intersect(bf$batch.id2, sample.id.validate))
+
+bf <- bulk[bulk$library_type=="RiboZeroGold",]
+length(intersect(bf$batch.id2, sample.id.train)) 
+length(intersect(bf$batch.id2, sample.id.validate))
 ```
 
 snRNA-seq
 
 ```{r}
-sce <- combineCols(mae[["snrnaseq.k2.all"]], mae[["snrnaseq.k2.all"]])
+sce <- rbind(mae[["snrnaseq.k2.all"]], mae[["snrnaseq.k2.all"]])
 
 # counts summary
 assays(sce)[["counts"]] |> colSums() |> summary()
@@ -128,6 +155,10 @@ apply(assays(bulk)[["counts"]],2,function(ci){length(ci[ci==0])}) |> summary()
 RNAscope
 
 ```{r}
+img <- mae[["sce.img"]]
+length(unique(img$Sample))
+length(intersect(img$Sample, sample.id.train))
+length(intersect(img$Sample, sample.id.validate))
 ```
 
 # Session info

cohort1/notebooks/05_bulk/01_plot_manual_save.Rmd

@@ -76,7 +76,7 @@ Proportions scatterplots.
 new.plot.scatter <- 
   ggplot(dfp, aes(x = true.neuron, y = neuron, color = bulk.scale.type)) + 
   theme_bw() + geom_point(alpha = 0.5) + geom_abline(slope = 1, intercept = 0) + 
-  facet_wrap(~experiment.type*assay.name.lutearg*s.set.label) +
+  facet_grid(experiment.type~assay.name.lutearg+s.set.label) +
   xlab("Known") + ylab("Predicted") + xlim(0, 1) + ylim(0, 1) +
   theme(axis.text.x = element_text(angle = 45, hjust = 1))
 new.plot.scatter
@@ -89,7 +89,7 @@ setwd("..")
 setwd("..")
 
 new.plot.path <- "./figures/05_bulk/figs3a_scatter.jpg"
-jpeg(new.plot.path, width = 7, height = 7, res = 400, units = "in")
+jpeg(new.plot.path, width = 8, height = 4.2, res = 400, units = "in")
 new.plot.scatter
 dev.off()
 ```

cohort1/notebooks/05_bulk/02_crossvalidate_save.Rmd

@@ -45,7 +45,7 @@ dfp <- na.omit(dfp)
 new.plot.scatter <- ggplot(dfp, aes(x = true.neuron, y = neuron)) + 
   theme_bw() + geom_point(alpha = 0.5) + 
   geom_abline(slope = 1, intercept = 0) + 
-  facet_wrap(~s.set.label*experiment.type) + xlab("Known") + 
+  facet_grid(s.set.label~experiment.type) + xlab("Known") + 
   ylab("Predicted") + xlim(0, 1) + ylim(0, 1) +
   theme(axis.text.x = element_text(angle = 45, hjust = 1))
 

cohort1/notebooks/07_adjustment/03_run_sopt_realbulk_all.Rmd

@@ -128,7 +128,7 @@ plotsList <- function(dfp, outlier = TRUE){
     newPlot <- ggplot(dfp, aes(x = true, y = value)) + 
       geom_point(size = 4, alpha = 0.5) + 
       geom_abline(slope = 1, intercept = 0) + 
-      facet_wrap(~cell.type*scale, nrow = 2) + xlab("Known") + 
+      facet_grid(cell.type~scale) + xlab("Known") + 
       ylab("Predicted") + theme_bw() + xlim(0, 1) + ylim(0, 1) +
         theme(axis.text.x = element_text(angle = 45, hjust = 1))
         
@@ -169,6 +169,8 @@ plotsListBISQUE <- plotsList(dfp.tall[dfp.tall$algorithm=="bisque",], outlier =
 Save NNLS multipanel plot.
 
 ```{r}
+setwd("..")
+setwd("..")
 
 jpeg("./figures/07_adjustment/fig3_bulk-results_nnls.jpg", 
      width = 6, height = 4, units = "in", res = 400)
@@ -183,7 +185,8 @@ dev.off()
 Save Bisque and MuSiC multipanel plot.
 
 ```{r}
-
+setwd("..")
+setwd("..")
 jpeg("./figures/07_adjustment/fig3_bulk-results_music-bisque.jpg", 
      width = 5.5, height = 8, units = "in", res = 400)
 grid.arrange(plotsListMUSIC[[1]] + ggtitle("MuSiC"), plotsListMUSIC[[2]] + ggtitle(""),