jokergoo
commented
2 years ago I think it is because you have too many columns. In the following code:
alter_fun = list(
background = function(x, y, w, h) {
grid.rect(x, y, w-unit(0.5, "mm"), h-unit(0.5, "mm"),
gp = gpar(fill = "beige", col = NA))
},
"wt" = function(x, y, w, h)
grid.rect(x, y, w * 0.9, h * 0.9,
gp = gpar(fill = alts_col["wt"], col = NA)),
...the background and other graphics actually have 0.5mm (or 0.05% of the cell width) white borders. When you save it as a png or view it not in its 100% size, these white borders are merged into neighbour pixels, which results in white lines (depends on neighbour colors or randomness).
To solve it, just remove these white borders:
alter_fun = list(
background = function(x, y, w, h) {
grid.rect(x, y, w, h,
gp = gpar(fill = "beige", col = NA))
},
"wt" = function(x, y, w, h)
grid.rect(x, y, w, h*0.9,
gp = gpar(fill = alts_col["wt"], col = NA)),
...
prinkessbambi
Hello, I've organized my oncoprint so that samples with alterations in gene1, 13, and 14 are displayed first/clustered. The oncoprint has these blocks of stripes throughout the rows corresponding to gene 1,13, and 14 that should not be there. I've ordered the matrix by my column order and inspected it-- my ordering is correct. These white lines don't represent samples with wt/missing data-- they should be colored in, as their corresponding samples in the matrix are marked as missense.
The matrix and column order are attached as CSV's because it is difficult to create a toy dataframe to show what's wrong.
test_mat.csv
col_order.csv
`
The alter fun function
alter_fun = list( background = function(x, y, w, h) { grid.rect(x, y, w-unit(0.5, "mm"), h-unit(0.5, "mm"), gp = gpar(fill = "beige", col = NA)) },
"wt" = function(x, y, w, h) grid.rect(x, y, w 0.9, h 0.9, gp = gpar(fill = alts_col["wt"], col = NA)),
"missense" = function(x, y, w, h) grid.rect(x, y, w 0.9, h 0.9, gp = gpar(fill = alts_col["missense"], col = NA)),
"damaging_mutation" = function(x, y, w, h) grid.rect(x, y, w 0.9, h 0.9, gp = gpar(fill = alts_col["damaging_mutation"], col = NA)),
"not_covered" = function(x, y, w, h) grid.rect(x, y, w 0.9, h 0.9, gp = gpar(fill = alts_col["not_covered"], col = NA)),
"rearrangement" = function(x, y, w, h) grid.rect(x, y, w 0.9, h 0.4, gp = gpar(fill = alts_col["rearrangement"], col = NA)),
"focal_gain" = function(x, y, w, h) grid.rect(x, y, w 0.9, h 0.9, gp = gpar(fill = alts_col["focal_gain"], col = NA)), "focal_loss" = function(x, y, w, h) grid.rect(x, y, w 0.9, h 0.9, gp = gpar(fill = alts_col["focal_loss"], col = NA)),
"arm_level_loss" = function(x, y, w, h) grid.rect(x, y, w 0.9, h 0.9, gp = gpar(fill = alts_col["arm_level_loss"], col = NA)),
"arm_level_gain" = function(x, y, w, h) grid.rect(x, y, w 0.9, h 0.9, gp = gpar(fill = alts_col["arm_level_gain"], col = NA)) )
For alterations on heatmap legend
alts_col <- c("blue", "red", "green", "skyblue", "orange", "beige", "white", "red", "blue") names(alts_col) <- c("focal_loss", "focal_gain", "rearrangement", "missense", "damaging_mutation", "wt", "not_covered", "arm_level_gain","arm_level_loss")
Read in data
umat <- read.csv("test_mat.csv", header = T, row.names = 1) ro <- read.csv("col_order.csv")
file = "test_mat.png" png( file , width = 1600, height = 1200); onc <- oncoPrint( umat, use_raster = F, column_order = ro$samples, show_pct = T, pct_side = "right", pct_gp = gpar(fontsize = 20), row_names_side = "left", row_names_gp = gpar(fontsize = 25), alter_fun_is_vectorized = TRUE, alter_fun = alter_fun, col = alts_col, heatmap_legend_param = list( labels_gp = gpar(fontsize = 20), title_gp = gpar(fontsize = 20, fontface = "bold"), title = "Alterations" ), right_annotation = rowAnnotation( #move barplot to left side rbar = anno_oncoprint_barplot( axis_param = list(direction = "reverse") )), row_order = rownames(umat), ); ComplexHeatmap::draw(onc, merge_legend = T); dev.off()
`