RScripts/Depression_Dataset_Analyses.R at main · amrithah93/RScripts · GitHub

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
attach(depression_cov_dx_symp_rest_conc_matched_qc) #modified dep dataset with first row removed to account for first missing subject in initial ICA analyses
library(corrplot)

#run correlation matrix to see which variables are related

install.packages("ggplot2")
install.packages("reshape2")
library(ggplot2)
library(reshape2)

#set up correlation
#omit missing data and create dataset for GSC items 1 and 2
#you have to use cbind else it won't work
#repeat process for multiple data sets to see which variables correlate

#test a GLM model with GSC_A_49
Y1 <- Depression_Loadings #list all loading parameters for patients
Y1 <-data.matrix(Y1)
model1 <- aov(Y1 ~ age + sex_coded+ GSC_A_49_Having_a_marked_decreased_interest_in_important_activities, data=depression_cov_dx_symp_rest_conc_matched_qc, na.action=na.omit)
summary.aov(model1) #view model output

#now test all 9 GSC models
Y2 <-Depression_Loadings
Y2 <-data.matrix(Y2)
model2 <-aov(Y2 ~ age + sex_coded + GSC_A_1_Feeling_depressed_or_being_in_a_sad_mood +
               GSC_A_2_Having_a_decreased_interest_in_things_that_are_usually_fun_including_sex +
               GSC_A_3_Experiencing_a_significant_change_in_weight_or_appetite_increased_or_decreased+
               GSC_A_4_Having_recurrent_thoughts_of_death_or_suicide + GSC_A_5_Experiencing_sleep_changes_such_as_a_lack_of_sleep_or_a_marked_increase_in_sleep +
               GSC_A_6_Feeling_physically_agitated_or_being_slowed_down + GSC_A_7_Having_feelings_of_low_energy_or_tiredness+
               GSC_A_8_Having_feelings_of_worthlessness_helplessness_hopelessness_or_guilt + GSC_A_9_Experiencing_decreased_concentration_or_memory)
summary.aov(model2)

#test models separately


# Define combinations of variables to analyze for each correlation pair
#GSC_A has 9 items, so most combos is 36 x 36
correlation_combo_list <-list(
  c("GSC_A_1_Feeling_depressed_or_being_in_a_sad_mood", "GSC_A_2_Having_a_decreased_interest_in_things_that_are_usually_fun_including_sex"),
  c("GSC_A_2_Having_a_decreased_interest_in_things_that_are_usually_fun_including_sex", "GSC_A_3_Experiencing_a_significant_change_in_weight_or_appetite_increased_or_decreased"),
  c("GSC_A_3_Experiencing_a_significant_change_in_weight_or_appetite_increased_or_decreased", "GSC_A_4_Having_recurrent_thoughts_of_death_or_suicide"),
  c("GSC_A_4_Having_recurrent_thoughts_of_death_or_suicide", "GSC_A_5_Experiencing_sleep_changes_such_as_a_lack_of_sleep_or_a_marked_increase_in_sleep"),
  c("GSC_A_5_Experiencing_sleep_changes_such_as_a_lack_of_sleep_or_a_marked_increase_in_sleep", "GSC_A_6_Feeling_physically_agitated_or_being_slowed_down"),
  c("GSC_A_6_Feeling_physically_agitated_or_being_slowed_down", "GSC_A_7_Having_feelings_of_low_energy_or_tiredness"),
  c("GSC_A_7_Having_feelings_of_low_energy_or_tiredness", "GSC_A_8_Having_feelings_of_worthlessness_helplessness_hopelessness_or_guilt"),
  c("GSC_A_8_Having_feelings_of_worthlessness_helplessness_hopelessness_or_guilt", "GSC_A_9_Experiencing_decreased_concentration_or_memory"))


#rename variables for readability
GSC1 <-GSC_A_1_Feeling_depressed_or_being_in_a_sad_mood
GSC2 <-GSC_A_2_Having_a_decreased_interest_in_things_that_are_usually_fun_including_sex
GSC3 <-GSC_A_3_Experiencing_a_significant_change_in_weight_or_appetite_increased_or_decreased
GSC4 <-GSC_A_4_Having_recurrent_thoughts_of_death_or_suicide
GSC5 <-GSC_A_5_Experiencing_sleep_changes_such_as_a_lack_of_sleep_or_a_marked_increase_in_sleep
GSC6 <-GSC_A_6_Feeling_physically_agitated_or_being_slowed_down
GSC7<-GSC_A_7_Having_feelings_of_low_energy_or_tiredness
GSC8<-GSC_A_8_Having_feelings_of_worthlessness_helplessness_hopelessness_or_guilt
GSC9 <-GSC_A_9_Experiencing_decreased_concentration_or_memory

#create data frame
df <-cbind.data.frame(GSC1, GSC2, GSC3, GSC4, GSC5, GSC6, GSC7, GSC8, GSC9)

# Create a data frame for each pair and omit NAs
cor1 <-na.omit(cbind(GSC1, GSC2))
cor2 <-na.omit(cbind(GSC2,GSC3))
cor3 <-na.omit(cbind(GSC3,GSC4))
cor4 <-na.omit(cbind(GSC4,GSC5))
cor5 <-na.omit(cbind(GSC5, GSC6))
cor6 <-na.omit(cbind(GSC6,GSC7))
cor7 <-na.omit(cbind(GSC7,GSC8))
cor8 <-na.omit(cbind(GSC8,GSC9))
full_dataset <-na.omit(cbind(df))

# Calculate the correlation coefficient
correlations_pair_1<- cor(cor1, method = c("pearson")) #remember to remove nas or it won't work
correlations_pair_2<- cor(cor2, method = c("pearson")) #remember to remove nas or it won't work
correlations_pair_3<- cor(cor3, method = c("pearson")) #remember to remove nas or it won't work
correlations_pair_4<- cor(cor4, method = c("pearson")) #remember to remove nas or it won't work
correlations_pair_5<- cor(cor5, method = c("pearson")) #remember to remove nas or it won't work
correlations_pair_6<- cor(cor6, method = c("pearson")) #remember to remove nas or it won't work
correlations_pair_7<- cor(cor7, method = c("pearson")) #remember to remove nas or it won't work
correlations_pair_8<- cor(cor8, method = c("pearson")) #remember to remove nas or it won't work
corr_full <-cor(full_dataset, method = c("pearson")) #full dataset

#correlation plots for all 8 pairs separately
ggcorrplot(cor(cor1)) #correlation matrix 1; make sure to zoom into figure to see values
ggcorrplot(correlations_pair_1,
           hc.order = TRUE,
           type = "full",
           lab = TRUE)

ggcorrplot(cor(cor2)) #correlation matrix 2; make sure to zoom into figure to see values
ggcorrplot(correlations_pair_2,
           hc.order = TRUE,
           type = "full",
           lab = TRUE)

ggcorrplot(cor(cor3)) #correlation matrix 3; make sure to zoom into figure to see values
ggcorrplot(correlations_pair_2,
           hc.order = TRUE,
           type = "full",
           lab = TRUE)

ggcorrplot(cor(cor4)) #correlation matrix 4; make sure to zoom into figure to see values
ggcorrplot(correlations_pair_4,
           hc.order = TRUE,
           type = "full",
           lab = TRUE)

ggcorrplot(cor(cor5)) #correlation matrix 5; make sure to zoom into figure to see values
ggcorrplot(correlations_pair_5,
           hc.order = TRUE,
           type = "full",
           lab = TRUE)

ggcorrplot(cor(cor6)) #correlation matrix 6; make sure to zoom into figure to see values
ggcorrplot(correlations_pair_6,
           hc.order = TRUE,
           type = "full",
           lab = TRUE)


ggcorrplot(cor(cor7)) #correlation matrix 7; make sure to zoom into figure to see values
ggcorrplot(correlations_pair_7,
           hc.order = TRUE,
           type = "full",
           lab = TRUE)


ggcorrplot(cor(cor8)) #correlation matrix 8; make sure to zoom into1 figure to see values
ggcorrplot(correlations_pair_8,
           hc.order = TRUE,
           type = "full",
           lab = TRUE)

#correlation plot of all together
ggcorrplot(corr_full) #correlation matrix full
ggcorrplot(corr_full,
           hc.order = TRUE,
           type = "full",
           lab = TRUE)

#do some preliminary ML classification
install.packages("OneR")