Mercurial > repos > ecology > ecology_stat_presence_abs

annotate graph_stat_presence_abs.r @ 1:9a2e0195bb43 draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
"planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 60627aba07951226c8fd6bb3115be4bd118edd4e"
author ecology
date Fri, 13 Aug 2021 18:18:21 +0000
parents e5552099d0e3
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
1 #Rscript
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
2
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
3 ################################
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
4 ## Median and dispersion ##
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
5 ################################
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
6
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
7 #####Packages : Cowplot
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
8 # ggplot2
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
9
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
10 #####Load arguments
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
11
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
12 args <- commandArgs(trailingOnly = TRUE)
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
13
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
14 if (length(args) == 0) {
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
15 stop("This tool needs at least one argument")
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
16 }else{
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
17 table <- args[1]
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
18 hr <- args[2]
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
19 var <- as.numeric(args[3])
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
20 spe <- as.numeric(args[4])
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
21 loc <- as.numeric(args[5])
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
22 time <- as.numeric(args[6])
1
9a2e0195bb43 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 60627aba07951226c8fd6bb3115be4bd118edd4e"
ecology
parents: 0
diff changeset
23 source(args[7])
0
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
24 }
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
25
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
26 if (hr == "false") {
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
27 hr <- FALSE
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
28 }else{
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
29 hr <- TRUE
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
30 }
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
31
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
32 #####Import data
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
33 data <- read.table(table, sep = "\t", dec = ".", header = hr, fill = TRUE, encoding = "UTF-8")
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
34 data <- na.omit(data)
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
35 colvar <- colnames(data)[var]
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
36 colspe <- colnames(data)[spe]
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
37 colloc <- colnames(data)[loc]
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
38 coltime <- colnames(data)[time]
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
39
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
40 data <- data[grep("^$", data[, spe], invert = TRUE), ]
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
41
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
42 #####Your analysis
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
43
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
44 ####Median and data dispersion####
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
45
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
46 #Median
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
47 graph_median <- function(data, var) {
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
48 graph_median <- ggplot2::ggplot(data, ggplot2::aes_string(y = var)) +
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
49 ggplot2::geom_boxplot(color = "darkblue") +
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
50 ggplot2::theme(legend.position = "none") + ggplot2::ggtitle("Median")
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
51
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
52 return(graph_median)
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
53
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
54 }
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
55
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
56 #Dispersion
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
57 dispersion <- function(data, var, var2) {
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
58 graph_dispersion <- ggplot2::ggplot(data) +
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
59 ggplot2::geom_point(ggplot2::aes_string(x = var2, y = var, color = var2)) +
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
60 ggplot2::scale_fill_brewer(palette = "Set3") +
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
61 ggplot2::theme(legend.position = "none", axis.text.x = ggplot2::element_text(angle = 90, vjust = 0.5, hjust = 1), plot.title = ggplot2::element_text(color = "black", size = 12, face = "bold")) + ggplot2::ggtitle("Dispersion")
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
62
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
63 return(graph_dispersion)
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
64
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
65 }
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
66
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
67 #The 2 graph
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
68 med_disp <- function(med, disp) {
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
69 graph <- cowplot::plot_grid(med, disp, ncol = 1, nrow = 2, vjust = -5, scales = "free")
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
70
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
71 ggplot2::ggsave("Med_Disp.png", graph, width = 12, height = 20, units = "cm")
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
72 }
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
73
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
74
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
75 #### Zero problem in data ####
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
76
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
77 #Put data in form
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
78
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
79 data_num <- make_table_analyse(data, colvar, colspe, colloc, coltime)
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
80 nb_spe <- length(unique(data[, spe]))
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
81 nb_col <- ncol(data_num) - nb_spe + 1
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
82 data_num <- data_num[, nb_col:ncol(data_num)]
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
83
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
84 #Presence of zeros in the data
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
85 mat_corr <- function(data) {
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
86 cor(data)
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
87 }
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
88 p_mat <- function(data) {
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
89 ggcorrplot::cor_pmat(data)
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
90 } # compute a matrix of correlation p-values
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
91
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
92 graph_corr <- function(data_num) {
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
93 graph <- ggcorrplot::ggcorrplot(mat_corr(data_num), method = "circle", p.mat = p_mat(data_num), #barring the no significant coefficient
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
94 ggtheme = ggplot2::theme_gray, colors = c("#00AFBB", "#E7B800", "#FC4E07"))
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
95
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
96 ggplot2::ggsave("0_pb.png", graph)
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
97 }
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
98
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
99 ##Med and disp
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
100 med <- graph_median(data, var = colvar)
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
101 disp <- dispersion(data, var = colvar, var2 = colspe)
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
102 med_disp(med = med, disp = disp)
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
103
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
104 ##O problem
e5552099d0e3 "planemo upload for repository https://github.com/Marie59/Data_explo_tools commit 2f883743403105d9cac6d267496d985100da3958"
ecology
parents:
diff changeset
105 graph_corr(data_num)