-
Notifications
You must be signed in to change notification settings - Fork 1
/
QUESTIONNAIRE_2_cosine_similarity.R
200 lines (148 loc) · 7 KB
/
QUESTIONNAIRE_2_cosine_similarity.R
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
setwd("data/")
ultra_lals <- c("03", "04", "07", "08", "13", "15", "16", "20", "25", "29", "33", "45", "46", "47", "49", "58")
ultra_lals <- paste0("LAL-00", ultra_lals)
# Load numeric table
df <- read.table("OUTPUT_lookalike_form_numeric.csv",sep = ",", stringsAsFactors = F, header = T)
df <- df[df$pair_id %in% ultra_lals,]
# I remove physical traits and of place of origin
categoric_physical_or_origin_traits <- c("Country",
"Gender",
"Language",
"Race",
"Current.Recidence.Category",
"Hair.Color",
"Hair.Shape",
"Eyes.Color",
"Country_born",
"City_village_born",
"Country_residence",
"City_village_residence",
"Lived_elsewhere")
# Remove categoric variables (all related to origin or physical traits)
df <- df[, !colnames(df) %in% categoric_physical_or_origin_traits]
# Remove columns with all 0
all_zeros_cols <- colnames(df)[3:ncol(df)][apply(df[,3:ncol(df)], 2, sum) == 0]
df <- df[, !colnames(df) %in% all_zeros_cols]
num_m <- df[, 3:ncol(df)]
rownames(num_m) <- df$ID_PEBC
# Normalize columns to 1
for(col_x in 3:ncol(df)) {
df[, col_x] <- df[, col_x]/max(df[, col_x], na.rm = T)
}
# order matrix
df <- df[order(df$ID_PEBC), ]
norm1_df <- df
rm(df)
cosine_sim <- function(a, b) crossprod(a,b)/sqrt(crossprod(a)*crossprod(b))
cosine_matrix <- matrix(as.numeric(NA), ncol = nrow(norm1_df), nrow = nrow(norm1_df))
for(x in 1:nrow(norm1_df)) {
for(y in 1:nrow(norm1_df)) {
a <- as.numeric(norm1_df[x, 3:ncol(norm1_df)])
b <- as.numeric(norm1_df[y, 3:ncol(norm1_df)])
cosine_matrix[x, y] <- cosine_sim(a,b)
}
}
colnames(cosine_matrix) <- norm1_df$ID_PEBC
rownames(cosine_matrix) <- norm1_df$ID_PEBC
library(gplots)
heatmap.2(cosine_matrix,
Rowv = F,
Colv = F,
dendrogram = "none",
colsep = 1:ncol(cosine_matrix), rowsep = 1:nrow(cosine_matrix),
density.info = "none",
cexCol = 0.8,
cexRow = 0.8,
col = colorRampPalette(c("white", "lightblue", "darkblue"))(80)[5:80],
sepwidth = c(0.01, 0.01),
margins = c(10,10),
sepcolor = "white", trace = "none")
# Ranking of scores
ranking_scores <- c()
for(lal in unique(norm1_df$pair)) {
lal_A <- paste0(lal, "-A")
lal_B <- paste0(lal, "-B")
ranking_scores <- c(ranking_scores, cosine_matrix[lal_A, lal_B])
names(ranking_scores)[length(ranking_scores)] <- lal
}
ranking_scores <- sort(ranking_scores, decreasing = T)
ranking_scores <- matrix(round(ranking_scores, 4), ncol = 1, dimnames = list(names(ranking_scores), "cosine_score"))
ultraLAL <- paste0("LAL-00", c("07", "13", "15", "16", "25", "33", "45", "47", "58"))
distance_score <- cbind(ranking_scores, order = order(ranking_scores[, "cosine_score"], decreasing = T))
distance_score <- cbind(distance_score, ultraLAL = rownames(distance_score) %in% ultraLAL)
distance_score[, "ultraLAL"] <- ifelse(distance_score[, "ultraLAL"] == 1, yes = "*", no = "")
distance_score <- cbind(rownames(distance_score), distance_score)
# write.table(distance_score, file = "cosine_ranking_with_age_height_weight_ULTRALALs_only.csv", quote = F, sep = ",", row.names = F)
# write.table(cosine_matrix , file = "cosine_matrix_with_age_height_weight_ULTRALALs_only.csv", quote = F, sep = ",", row.names = T)
save(norm1_df, file = "lookalike_form_normalized_to_1.RData")
cosine_matrix <- matrix(as.numeric(NA), ncol = nrow(norm1_df), nrow = nrow(norm1_df))
for(x in 1:nrow(norm1_df)) {
for(y in 1:nrow(norm1_df)) {
a <- as.numeric(norm1_df[x, 3:ncol(norm1_df)])
b <- as.numeric(norm1_df[y, 3:ncol(norm1_df)])
cosine_matrix[x, y] <- cosine_sim(a,b)
}
}
colnames(cosine_matrix) <- norm1_df$ID_PEBC
rownames(cosine_matrix) <- norm1_df$ID_PEBC
library(gplots)
# pdf("cosine_matrix_filtering_out_age_height_weight_ULTRALALs_only.pdf", height = 12, width = 12)
heatmap.2(cosine_matrix,
Rowv = F,
Colv = F,
dendrogram = "none",
colsep = 1:ncol(cosine_matrix), rowsep = 1:nrow(cosine_matrix),
density.info = "none",
cexCol = 0.7,
cexRow = 0.7,
col = colorRampPalette(c("white", "lightblue", "darkblue"))(80)[5:80],
sepwidth = c(0.01, 0.01),
margins = c(10,10),
sepcolor = "white", trace = "none")
# dev.off()
# Ranking of scores
ranking_scores <- c()
for(lal in unique(norm1_df$pair)) {
lal_A <- paste0(lal, "-A")
lal_B <- paste0(lal, "-B")
ranking_scores <- c(ranking_scores, cosine_matrix[lal_A, lal_B])
names(ranking_scores)[length(ranking_scores)] <- lal
}
ranking_scores <- sort(ranking_scores, decreasing = T)
ranking_scores <- matrix(round(ranking_scores, 4), ncol = 1, dimnames = list(names(ranking_scores), "cosine_score"))
ultraLAL <- paste0("LAL-00", c("07", "13", "15", "16", "25", "33", "45", "47", "58"))
distance_score <- cbind(ranking_scores, order = order(ranking_scores[, "cosine_score"], decreasing = T))
distance_score <- cbind(distance_score, ultraLAL = rownames(distance_score) %in% ultraLAL)
distance_score[, "ultraLAL"] <- ifelse(distance_score[, "ultraLAL"] == 1, yes = "*", no = "")
distance_score <- cbind(rownames(distance_score), distance_score)
# write.table(distance_score, file = "cosine_ranking_no_age_height_weight_ULTRALALs_only.csv", quote = F, sep = ",", row.names = F)
# write.table(cosine_matrix, file = "cosine_matrix_no_age_height_weight_ULTRALALs_only.csv", quote = F, sep = ",", row.names = T)
### Are samples more similar to their than with the others?
cosine_matrix_extra <- cosine_matrix
for(lal in unique(norm1_df$pair)) {
lal_A <- paste0(lal, "-A")
lal_B <- paste0(lal, "-B")
cosine_matrix_extra[lal_A, lal_A] <- NA
cosine_matrix_extra[lal_B, lal_B] <- NA
cosine_matrix_extra[lal_A, lal_B] <- NA
cosine_matrix_extra[lal_B, lal_A] <- NA
}
for(x in 3:nrow(cosine_matrix_extra)) {
for(y in 1:x) {
cosine_matrix_extra[x, y] <- NA
}
}
# intra and extra scores
intra_pair <- c(ranking_scores)
extra_pair <- c(cosine_matrix_extra[!is.na(cosine_matrix_extra)])
# pdf("intra_extra_pair_cosine_similarity_boxplot.pdf", width = 5, height = 5)
bp <- boxplot(list(intra_pair, extra_pair),
main = "Cosine similarity matrix",
ylab = "Cosine distance",
ylim = c(0.2,1), lwd = 1.25, col = "gray70")
arrows(1,0.95,2, 0.95, code = 3, angle = 90, length = 0.07, lwd = 1.5)
# label <- ifelse(t.test(intra_pair, extra_pair)$p.value < 0.05, yes = "*", no = "n.s.")
# text(x = 1.5, y = 0.98, labels = label, cex = 1.5)
pval <- t.test(intra_pair, extra_pair)$p.value
text(x = 1.5, y = 0.98, labels = paste0("P = ", round(pval, 5)), cex = 1)
# dev.off()