tanu/LSHTM_analysis
1
0
Fork 0
Code Issues Pull requests Releases Wiki Activity

config: add tile stuff for all targets. Other functions: many rewrites!

Browse source
This commit is contained in:
Tanushree Tunstall 2022-08-15 16:29:38 +01:00
parent a3e5283a9b
commit 0cc7a8fcae
11 changed files with 163 additions and 200 deletions
Download patch file Download diff file Expand all files Collapse all files

2
config/alr.R
View file

@ -107,4 +107,6 @@ if ( all(c1, c2, c3) ) {
aa_pos_lig1 = NULL
aa_pos_lig2 = NULL
aa_pos_lig3 = NULL
tile_map=data.frame(tile=c("ALR","DSL","CDL","Ca"),
tile_colour=c("green","darkslategrey","navyblue","purple"))

4
config/embb.R
View file

@ -116,3 +116,7 @@ cat("\n==================================================="
aa_pos_lig1 = aa_pos_dsl #slategray
aa_pos_lig2 = aa_pos_cdl #navy blue
aa_pos_lig3 = aa_pos_ca #purple
tile_map=data.frame(tile=c("EMB","DSL","CDL","Ca"),
tile_colour=c("green","darkslategrey","navyblue","purple"))

3
config/gid.R
View file

@ -132,3 +132,6 @@ cat("\n==================================================="
aa_pos_lig1 = aa_pos_rna
aa_pos_lig2 = aa_pos_sam
aa_pos_lig3 = aa_pos_amp
tile_map=data.frame(tile=c("GID","DSL","CDL","Ca"),
tile_colour=c("green","darkslategrey","navyblue","purple"))

3
config/katg.R
View file

@ -106,3 +106,6 @@ cat("\n==================================================="
aa_pos_lig1 = aa_pos_hem
aa_pos_lig2 = NULL
aa_pos_lig3 = NULL
tile_map=data.frame(tile=c("KAT","DSL","CDL","Ca"),
tile_colour=c("green","darkslategrey","navyblue","purple"))

3
config/pnca.R
View file

@ -54,3 +54,6 @@ cat("\n==================================================="
aa_pos_lig1 = aa_pos_fe
#aa_pos_lig2 = aa_pos_catalytic
#aa_pos_lig3 = aa_pos_hbond
tile_map=data.frame(tile=c("PZA","DSL","CDL","Ca"),
tile_colour=c("green","darkslategrey","navyblue","purple"))

3
config/rpob.R
View file

@ -59,3 +59,6 @@ cat("\n==================================================="
aa_pos_lig1 = NULL
aa_pos_lig2 = NULL
aa_pos_lig3 = NULL
tile_map=data.frame(tile=c("RPO","DSL","CDL","Ca"),
tile_colour=c("green","darkslategrey","navyblue","purple"))

47
scripts/functions/bp_subcolours.R
View file

@ -60,17 +60,9 @@ bp_stability_hmap <- function(plot_df = merged_df3
)
{
################################################
# Custom 2: x-axis geom tiles ~ lig distance
################################################
#=========================
# Build data with colours
# ~ ligand distance
#=========================
# plot_df = generate_distance_colour_map(plot_df, debug=TRUE)
# order the df by position and ensure it is a factor
# order the df by position and ensure it is a factor
plot_df = plot_df[order(plot_df[[xvar_colname]]), ]
plot_df[[xvar_colname]] = factor(plot_df[[xvar_colname]])
@ -79,26 +71,30 @@ bp_stability_hmap <- function(plot_df = merged_df3
# stability values isolated to help with generating column called: 'group'
my_grp = plot_df[[stability_colname]]
cat( "\nLength of nsSNPs:", length(my_grp)
, "\nLength of unique values for nsSNPs:", length(unique(my_grp)) )
# cat( "\nLength of nsSNPs:", length(my_grp)
# , "\nLength of unique values for nsSNPs:", length(unique(my_grp)) )
#
# Add col: 'group'
plot_df$group = paste0(plot_df[[stability_outcome_colname]], "_", my_grp, sep = "")
# check unique values in normalised data
cat("\nNo. of unique values in", stability_colname, "no rounding:"
, length(unique(plot_df[[stability_colname]])))
# cat("\nNo. of unique values in", stability_colname, "no rounding:"
# , length(unique(plot_df[[stability_colname]])))
# Call the function to create the palette based on the group defined above
#subcols_ps
subcols_bp_hmap = ColourPalleteMulti(plot_df, stability_outcome_colname, stability_colname)
cat("\nNo. of sub colours generated:", length(subcols_bp_hmap))
anno_bar=position_annotation(plot_df,
aa_pos_drug=aa_pos_drug,
active_aa_pos=active_aa_pos,
aa_pos_lig1=aa_pos_lig1,
aa_pos_lig2=aa_pos_lig2,
aa_pos_lig3=aa_pos_lig3
)
#-------------------------------
# Generate the subcols barplot
#-------------------------------
cowplot::plot_grid(
ggplot(plot_df, aes_string(x = xvar_colname
# , ordered = T)
@ -111,7 +107,8 @@ bp_stability_hmap <- function(plot_df = merged_df3
# scale_x_discrete("Position", labels=factor(plot_df$position)) +
theme(
axis.text.x = element_text(size = my_xaxls
panel.grid = element_line(color="lightgrey", size=0.125)
, axis.text.x = element_text(size = my_xaxls
, angle = 90
, hjust = 1
, vjust = 0.4)
@ -128,25 +125,17 @@ bp_stability_hmap <- function(plot_df = merged_df3
# , panel.grid = element_blank()
, panel.background = element_rect(fill = "transparent", colour=NA)
) +
labs(title = p_title
, x = my_xlab
, y = my_ylab),
NULL,
position_annotation(plot_df,
aa_pos_drug=aa_pos_drug,
active_aa_pos=active_aa_pos,
aa_pos_lig1=aa_pos_lig1,
aa_pos_lig2=aa_pos_lig2,
aa_pos_lig3=aa_pos_lig3
)
,
anno_bar,
#generate_distance_legend(plot_df),
ncol = 1,
align = "v",
rel_heights = c(10,-0.1,1)
rel_heights = c(6,-0.1,1)
#rel_widths = c(9/10, 0.4/10)
)
}
#bp_stability_hmap(merged_df3)
# bp_stability_hmap(small_df3)

25
scripts/functions/generate_distance_colour_map.R
View file

@ -1,9 +1,9 @@
# takes a dataframe and returns the same dataframe with two extra columns for colours and position
library('viridis')
generate_distance_colour_map = function(plot_df,
xvar_colname = "position",
lig_dist_colname = "ligand_distance",
#lig_dist_colours = c("green", "yellow", "orange", "red"),
lig_dist_colours = c("green", "yellow", "magenta"),
debug = TRUE
)
{
@ -16,12 +16,9 @@ generate_distance_colour_map = function(plot_df,
lig_min = min(round(plot_df[[lig_dist_colname]]), na.rm = T); lig_min
lig_max = max(round(plot_df[[lig_dist_colname]]), na.rm = T); lig_max
lig_mean = round(mean(round(plot_df[[lig_dist_colname]]), na.rm = T)); lig_mean
# Create mapping colour key
n_colours = length(sort(unique(round(plot_df[[lig_dist_colname]], digits = 0)))); n_colours
lig_cols = colorRampPalette(lig_dist_colours)(n_colours); lig_cols
ligD_valsR = sort(unique(round(plot_df[[lig_dist_colname]], digits = 0))); ligD_valsR
n_colours = length(sort(unique(round(plot_df[[lig_dist_colname]], digits = 0))))
lig_cols = magma(n_colours, direction=-1)
ligD_valsR = sort(unique(round(plot_df[[lig_dist_colname]], digits = 0)))
if (debug) {
length(ligD_valsR)
@ -43,15 +40,9 @@ generate_distance_colour_map = function(plot_df,
names(ligDcolKey)
cat("\nSuccessful: Mapping b/w", lig_dist_colname, "and colours")
}
#-------------------------------------
# merge colour key with plot_df
#--------------------------------------
plot_df = merge(plot_df, ligDcolKey, by = 'lig_distR')
# plot_df_check = as.data.frame(cbind(position = plot_df[[xvar_colname]]
# , ligD = plot_df[[lig_dist_colname]]
# , ligDR = plot_df$lig_distR
# , ligD_cols = plot_df$ligD_colours))
return(plot_df)
}
@ -75,9 +66,9 @@ generate_distance_legend = function(plot_df,
geom_tile(aes(fill = .data[[lig_dist_colname]])
, colour = "white") +
scale_fill_gradient2(midpoint = lig_mean
, low = "green"
, mid = "yellow"
, high = "magenta"
, low = magma(3, direction=-1)[1]
, mid = magma(3, direction=-1)[2]
, high = magma(3, direction=-1)[3]
, breaks = labels
, limits = c(lig_min, lig_max)
, labels = labelsD

77
scripts/functions/logoP_or.R
View file

@ -49,16 +49,16 @@ LogoPlotCustomH <- function(plot_df
, y_axis_increment = 50
, x_lab = "Position"
, y_lab = "Odds Ratio"
, x_ats = 12 # text size
, x_ats = 6 # text size
, x_tangle = 90 # text angle
, y_ats = 22
, y_ats = 11
, y_tangle = 0
, x_tts = 20 # title size
, y_tts = 23
, x_tts = 10 # title size
, y_tts = 11
, leg_pos = "none" # can be top, left, right and bottom or c(0.8, 0.9)
, leg_dir = "horizontal" #can be vertical or horizontal
, leg_ts = 15 # leg text size
, leg_tts = 16 # leg title size
, leg_ts = 7 # leg text size
, leg_tts = 8 # leg title size
, tpos0 = 0 # 0 is a magic number that does my sensible default
, tW0 = 1
, tH0 = 0.3
@ -68,16 +68,16 @@ LogoPlotCustomH <- function(plot_df
#################################
# Data processing for logo plot
#################################
plot_df = generate_distance_colour_map(plot_df, debug=TRUE)
unique_colour_map = unique(plot_df[,c("position","ligD_colours")])
unique_colour_map = unique_colour_map[order(unique_colour_map$position), ]
rownames(unique_colour_map) = unique_colour_map$position
unique_colour_map2 = unique_colour_map
unique_colour_map2$position=as.factor(unique_colour_map2$position)
unique_colour_map2$ligD_colours = as.factor(unique_colour_map2$ligD_colours)
# plot_df = generate_distance_colour_map(plot_df, debug=TRUE)
# unique_colour_map = unique(plot_df[,c("position","ligD_colours")])
# unique_colour_map = unique_colour_map[order(unique_colour_map$position), ]
# rownames(unique_colour_map) = unique_colour_map$position
#
# unique_colour_map2 = unique_colour_map
# unique_colour_map2$position=as.factor(unique_colour_map2$position)
# unique_colour_map2$ligD_colours = as.factor(unique_colour_map2$ligD_colours)
#
if (rm_empty_y){
@ -147,7 +147,7 @@ LogoPlotCustomH <- function(plot_df
cat('\nSelected colour scheme:', my_logo_col
, "\nUsing grey theme")
theme_bgc = "grey"
theme_bgc = "white"
xfont_bgc = "black"
yfont_bgc = "black"
xtt_col = "black"
@ -179,7 +179,7 @@ LogoPlotCustomH <- function(plot_df
, angle = x_tangle
, colour = xfont_bgc
, vjust = 0.4
, margin = margin(t=-8,r=0,b=0,l=0, unit="mm")
, margin = margin(t=0,r=0,b=0,l=0, unit="mm")
)
, axis.text.y = element_text(size = y_ats
, angle = y_tangle
@ -211,47 +211,6 @@ LogoPlotCustomH <- function(plot_df
#, labels = seq(0, (y_lim), by = y_axis_increment)
, limits = c(0, y_lim)) +
labs(y=y_lab),
NULL,
ggplot(data=unique_colour_map2, aes(
x=factor(position), 0 # heat-mapped distance tiles along the bot
, fill = position
, colour = position
, linetype = "blank"
)
) +
geom_tile() +
theme(
axis.text.x = element_blank()
# axis.text.x = element_text(size = x_ats
# , angle = x_tangle
# , hjust = 1
# , vjust = 0.4
# , colour = xfont_bgc)
, axis.text.y = element_blank()
, axis.ticks.y = element_blank()
, axis.ticks.x = element_blank()
, axis.title.x = element_blank()
, axis.title.y = element_text(size = y_tts
, colour = ytt_col)
, legend.title = element_text(size = leg_tts
, colour = ytt_col)
, legend.text = element_text(size = leg_ts)
, legend.position = leg_pos
, legend.direction = leg_dir
#, plot.background = element_blank()
, plot.margin = margin(t=0,b=1)
, panel.grid=element_blank()
, plot.background = element_rect(fill = theme_bgc, colour=NA)
, panel.background = element_rect(fill = "transparent", colour=NA)
#, panel.background = element_rect(fill = theme_bgc)
#, panel.background = element_blank()
) +
#scale_x_discrete(x_lab, labels=factor(unique_colour_map$position)) +
scale_color_manual(values=unique_colour_map$ligD_colours) +
scale_fill_manual(values=unique_colour_map$ligD_colours) +
labs(y = NULL),
NULL,
position_annotation(plot_df,
bg = theme_bgc,
aa_pos_drug=aa_pos_drug,
@ -261,7 +220,7 @@ LogoPlotCustomH <- function(plot_df
aa_pos_lig3=aa_pos_lig3
),
ncol=1, align='v', rel_heights = c(16,0,1,0,1)
ncol=1, align='v', rel_heights = c(6,1)
)
}
#LogoPlotCustomH(small_df3)

59
scripts/functions/logoP_snp.R
View file

@ -208,25 +208,20 @@ LogoPlotSnps <- function(plot_df
xtt_col = "black"
ytt_col = "black"
}
position_mt = as.numeric(colnames(tab_mt))
position_wt = as.numeric(colnames(tab_wt))
#####################################
# Generating logo plots for nsSNPs
#####################################
#-------------------
# Mutant logo plot
#-------------------
logo_top =ggseqlogo(tab_mt
, method = 'custom'
, col_scheme = my_logo_col
, seq_type = 'aa') +
scale_x_continuous(breaks = 1:ncol(tab_mt)
, expand = c(0.01,0)
, labels = colnames(tab_mt))+
scale_y_continuous(breaks = 0:(max_mult_mut-1)
, labels = c(1:max_mult_mut)
, limits = c(0, max_mult_mut)) +
logo_top = ggplot() +
geom_logo(tab_mt
, method = 'custom'
, col_scheme = my_logo_col
, seq_type = 'aa') +
theme_nothing() +
ylab(y_lab) +
theme(text=element_text(family="FreeSans")
, legend.position = leg_pos
@ -239,7 +234,9 @@ LogoPlotSnps <- function(plot_df
, angle = x_tangle
#, hjust = 1
#, vjust = 0.4
, colour = xfont_bgc)
, colour = xfont_bgc
#, margin = margin(t = 0.1)
)
, axis.text.y = element_text(size = y_ats
, angle = y_tangle
, hjust = 1
@ -249,17 +246,25 @@ LogoPlotSnps <- function(plot_df
# , colour = xtt_col)
, axis.title.x = element_blank()
, axis.title.y = element_text(size = y_tts
, angle = 90
, colour = ytt_col)
, plot.background = element_rect(fill = theme_bgc, colour=NA)
) +
scale_x_discrete("Position"
, labels = position_mt
, limits = factor(1:length(position_mt))
)
logo_bottom = ggseqlogo(tab_wt
, method = 'custom'
, col_scheme = my_logo_col
, seq_type = 'aa') +
scale_x_continuous(breaks = 1:ncol(tab_wt)
, expand = c(0.01,0)
, labels = as.factor(colnames(tab_wt))) +
logo_bottom = ggplot() +
geom_logo(tab_wt
, method = 'custom'
, col_scheme = my_logo_col
, seq_type = 'aa') +
theme_nothing() +
scale_x_discrete("Position"
, labels = x_axis_colname
, limits = factor(1:length(x_axis_colname))) +
theme(text = element_text(family="FreeSans")
, legend.position = "none"
, axis.text.x = element_blank()
@ -278,13 +283,13 @@ LogoPlotSnps <- function(plot_df
aa_pos_lig2=aa_pos_lig2,
aa_pos_lig3=aa_pos_lig3)
aligned=align_plots(logo_top, logo_bottom, anno_bar, align='v', axis='lr')
#aligned=align_plots(logo_top, logo_bottom, anno_bar, align='vh', axis='lr')
cowplot::plot_grid(
aligned[[1]], aligned[[2]], aligned[[3]],
logo_top, logo_bottom, anno_bar,
#aligned[[1]], aligned[[2]], aligned[[3]],
ncol=1,
#align = "v",
rel_heights = c(7, 1,1),
rel_widths = c(1,1,0.75)
align = "v",
rel_heights = c(7, 1,1)
)
# cowplot::plot_grid(
@ -303,4 +308,4 @@ LogoPlotSnps <- function(plot_df
#------------------
}
#LogoPlotSnps(mutable_df3)
#LogoPlotSnps(small_df3)

125
scripts/functions/position_annotation.R
View file

@ -49,77 +49,78 @@ position_annotation=function(plot_df,
plot_df = generate_distance_colour_map(plot_df, debug=TRUE)
cowplot::plot_grid(
ggplot(plot_df,
aes(x=factor(position), # THIS STUPID FUCKING FACTOR THING
)
heat_bar = ggplot(plot_df,
aes(x=factor(position)) # THIS STUPID FUCKING FACTOR THING
) +
geom_tile(aes(y=0),
fill=plot_df$ligD_colours) +
# scale_x_discrete("Position", labels=factor(plot_df$position)) +
theme_nothing() +
theme(#axis.text.x = element_text(angle = 90, size = 6),
title = element_blank()
) + # enable for alignment debug
labs(x = NULL, y = NULL) #end of distance-heat-bar
#NULL,
pos_tiles = ggplot(plot_df,
aes(x=factor(position), # THIS STUPID FUCKING FACTOR THING
)
) +
# geom_tile(aes(y = 0, fill = col_aa, colour = col_aa)
# ) +
geom_tile(aes(y = 1, fill = bg_all, colour = bg_all)
) +
geom_tile(aes(y=0),
fill=plot_df$ligD_colours) +
#scale_x_discrete("Position", labels=factor(plot_df$position)) +
#theme_nothing() +
theme(plot.background = element_rect(fill = "transparent", colour=NA),
plot.margin = margin(t=0,b=0),
axis.ticks.x = element_blank(),
axis.ticks.y = element_blank(),
axis.text.y = element_blank(),
panel.grid = element_blank(),
panel.background = element_rect(fill = "transparent", colour=NA),
) +
labs(x = NULL, y = NULL), #end of distance-heat-bar
#NULL,
ggplot(plot_df,
aes(x=factor(position), # THIS STUPID FUCKING FACTOR THING
#reorder(ligand_distance)
)
geom_tile(aes(y = 2, fill = col_bg1, colour = col_bg1)
) +
geom_tile(aes(y = 3, fill = col_bg2, colour = col_bg2)
) +
geom_tile(aes(y = 4, fill = col_bg3, colour = col_bg3)
) +
# geom_tile(aes(y = 0, fill = col_aa, colour = col_aa)
# ) +
geom_tile(aes(y = 1, fill = bg_all, colour = bg_all)
) +
geom_tile(aes(y = 2, fill = col_bg1, colour = col_bg1)
) +
geom_tile(aes(y = 3, fill = col_bg2, colour = col_bg2)
) +
geom_tile(aes(y = 4, fill = col_bg3, colour = col_bg3)
) +
#scale_x_discrete("Position", labels=factor(plot_df$position)) +
scale_color_manual(values = c(
"brown"="brown",
"green"="#00ff00",
"transparent"="transparent",
"slategrey"="#2f4f4f",
"navyblue"="#000080",
"purple"="#a020f0"
),
expand=c(0,0)
) +
scale_fill_manual(values = c(
"brown"="brown",
"green"="#00ff00",
"transparent"="transparent",
"slategrey"="#2f4f4f",
"navyblue"="#000080",
"purple"="#a020f0"
),
expand=c(0,0)
) +
#scale_x_continuous(expand=c(0,0)) +
#scale_y_continuous(expand=c(0,0)) +
theme_nothing() +
#scale_x_discrete("Position", labels=factor(plot_df$position)) +
scale_color_manual(values = c(
"brown"="brown",
"green"="#00ff00",
"transparent"="transparent",
"slategrey"="#2f4f4f",
"navyblue"="#000080",
"purple"="#a020f0"
),
#expand=c(0,0)
) +
scale_fill_manual(values = c(
"brown"="brown",
"green"="#00ff00",
"transparent"="transparent",
"slategrey"="#2f4f4f",
"navyblue"="#000080",
"purple"="#a020f0"
),
#expand=c(0,0)
) +
theme_nothing() +
theme(plot.background = element_rect(fill = bg, colour=NA),
#plot.margin = margin(t=0,b=0),
panel.background = element_rect(fill = bg, colour=NA),
legend.position = "none"
) +
labs(x = NULL, y= NULL)
theme(plot.background = element_rect(fill = bg, colour=NA),
plot.margin = margin(t=0,b=0)) +
labs(x = NULL, y = NULL),
heat_legend=get_legend(heat_bar)
out_plot=cowplot::plot_grid(
heat_bar,
NULL,
pos_tiles,
ncol=1,
align='v',
rel_heights = c(1,
#-0.1,
1)
-0.1,
2)
)
return(out_plot)
}
# position_annotation(merged_df3,
# position_annotation(small_df3,
# aa_pos_drug=aa_pos_drug,
# active_aa_pos=active_aa_pos,
# aa_pos_lig1=aa_pos_lig1,