added peptide intensities of STANDARDs (boxplot and heatmap); corrected data...

added peptide intensities of STANDARDs (boxplot and heatmap); corrected data plotting with scale_y_log10 and now use intensities+1 to make sure we do not lose the zero intensities; added export of top10 LFQs per sample

ms_workflow/ms_data_prep.R

@@ -151,7 +151,7 @@ init_stat %>% select(GROUP, unique_entry, biased_entry, scrap_prop_unique, scrap
     filter(direction) %>%
     mutate(GROUP = str_replace(GROUP, ".proteinGroups.txt", "")) %>%
     na.omit() %>%
-    ggplot(aes(feature, lfq_intensity, fill = feature)) +
+    ggplot(aes(feature, lfq_intensity+1, fill = feature)) +
         geom_boxplot() +
         # geom_violin() +
         scale_y_log10() +
@@ -172,7 +172,7 @@ init_stat %>% select(GROUP, unique_entry, biased_entry, scrap_prop_unique, scrap
     mutate(sample = str_replace(sample, "lfq_intensity_", "") %>% str_replace_all(., oldNames, newNames),
         GROUP = str_replace(GROUP, ".proteinGroups.txt", "") %>% str_replace_all(., oldNames, newNames)) %>%
     na.omit() %>%
-    ggplot(aes(feature, lfq_intensity, fill = feature)) +
+    ggplot(aes(feature, lfq_intensity+1, fill = feature)) +
     geom_boxplot() +
 # geom_violin() +
     scale_y_log10() +
@@ -182,7 +182,10 @@ init_stat %>% select(GROUP, unique_entry, biased_entry, scrap_prop_unique, scrap
     scale_fill_manual(values=c("coral3", "coral1", "cadetblue3", "cadetblue")) +
     facet_wrap(~GROUP + sample)
 
-
+#'
+#' <br><br>
+#'
+#' ### LFQ and raw intensities of STANDARDS on protein level
 rbind(init_stat %>% filter(str_detect(protein_ids, "STANDARD")) %>%
     select(starts_with("lfq_intensity")) %>%
     gather(sample, intensities) %>% na.omit() %>%
@@ -191,7 +194,33 @@ rbind(init_stat %>% filter(str_detect(protein_ids, "STANDARD")) %>%
     select(starts_with("intensity_")) %>%
     gather(sample, intensities) %>% na.omit() %>%
     mutate(sample = str_replace(sample, "lfq_intensity_", ""), feature = "raw intensities")
-) %>% ggplot(aes(feature, intensities)) + geom_boxplot() + ggtitle("Intensities of STANDARDs")
+) %>% ggplot(aes(feature, intensities)) + geom_boxplot()
+
+
+#'
+#' <br><br>
+#'
+#' ### LFQ and raw intensities of STANDARDS on peptide level
+pepFiles <- list.files(ms_data_folder, pattern = "peptides.txt", full = TRUE)
+
+std_info = pepFiles %>%
+    map(~ read_tsv(.x) %>% pretty_columns() %>%
+        select(proteins, sequence, matches("^lfq_intensity_"), matches("^intensity_")) %>%
+        filter(str_detect(proteins, "STANDARD"))) %>%
+    setNames(basename(pepFiles)) %>%
+    map_df(~ gather(.x, feature, intensity, contains("intensity")) %>%
+    separate(feature, c("type", "sample"), sep = "sity_"), .id = 'GROUP') %>%
+    mutate(sample = str_replace_all(sample, oldNames, newNames), type = str_replace(type, "inten", "intensity"))
+
+print("STANDARD intensities on peptide level are available for the following samples:")
+unique(std_info$sample)
+
+
+std_info %>% ggplot(aes(sample, intensity, fill = type)) + geom_boxplot() + coord_flip()
+
+std_info %>%
+    ggplot(aes(sample, sequence, fill = intensity)) + geom_tile() + facet_wrap(~type) +
+    theme(axis.text.x = element_text(angle = 45, hjust = 1))
 
 
 # init_stat %>% select(GROUP, protein_ids, unique_entry, biased_entry, scrap_prop_unique, scrap_prop_biased, matches("lfq_intensity")) %>%
@@ -372,7 +401,7 @@ grid.arrange(p1, p2, nrow = 1)
 protsData %>% ggplot(aes(intensity)) + geom_histogram(fill = "lightcyan4") + scale_x_log10() + ggtitle("LFQ intensities") + facet_wrap(~sample)
 
 
-protsData %>% ggplot(aes(sample, intensity)) + geom_boxplot(fill = "lightcyan4") + ggtitle("LFQ intensities") + scale_y_log10() + coord_flip()
+protsData %>% ggplot(aes(sample, intensity+1)) + geom_boxplot(fill = "lightcyan4") + ggtitle("LFQ intensities") + scale_y_log10() + coord_flip()
 
 ## when using linear scale, we discard 0s to avoid biases means
 protsData %>% filter(intensity>0) %>% ggplot(aes(sample, intensity)) +
@@ -407,6 +436,7 @@ if("identification_type" %in% colnames(heg_data) == FALSE){
     d3 <- d1 + geom_point(aes(order, intensity, shape = identification_type), fill = "black") + scale_shape_manual(values=c(20, 1)) + facet_wrap(~sample, scales="free", ncol = 2) + ggtitle("LFQ intensities of top 10 proteins per sample - different scale")
 }
 
+write_tsv(heg_data, path=add_prefix("top10_lfq.txt"))
 
 #+ fig.height=length(unique(heg_data$sample))/2+5, eval=length(unique(heg_data$sample))>0
 d2