Log In Sign Up
Models:
Daniel Galbraith
DanielG/
MMDRP
0
Downloads: 1
[c3b4f8]: / R / 01_Dose-Response_Data_Preparation.R

Download this file

182 lines (147 with data), 8.9 kB 

  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
# PharmacoGx.R

# BiocManager::install("PharmacoGx")

require(PharmacoGx)

require(data.table)

path = "/Users/ftaj/OneDrive - University of Toronto/Drug_Response/"

dir.create(paste0(path, "Data/DRP_Training_Data"))



depmap_samples <- fread(paste0(path, "Data/DRP_Training_Data/DepMap_20Q2_Line_Info.csv"))



depmap_samples$stripped_cell_line_name

PharmacoGx::availablePSets()



# CTRPv2 ====

??summarizeSensitivityProfiles

require(PharmacoGx)



ctrp2 <- downloadPSet("CTRPv2_2015")



ctrp2_aac <- summarizeSensitivityProfiles(object = ctrp2, sensitivity.measure = "aac_recomputed", summary.stat = "median", fill.missing = T,)



fwrite(as.data.table(ctrp2@drug, keep.rownames = T), "Data/DRP_Training_Data/CTRP_DRUG_INFO.csv")

drug_info <- ctrp2@drug[, c("gene_symbol_of_protein_target", "target_or_activity_of_compound", "cpd_status")]

drug_info[1:5,]

drug_info <- as.data.table(drug_info, keep.rownames = T)



drug_info[rn %like% "Paclitaxel"]

drug_info[gene_symbol_of_protein_target != "" & cpd_status == "clinical"]$rn

# ctrp2@drug[1:20, c("drugid", "cpd_name")]

# ctrp2_ic50 <- PharmacoGx::summarizeSensitivityProfiles(object = ctrp2, sensitivity.measure = "ic50_recomputed", summary.stat = "median", fill.missing = T)

# Convert to data.table

ctrp2_aac <- as.data.table(ctrp2_aac, keep.rownames = T)

# ctrp2_ic50 <- as.data.table(ctrp2_ic50, keep.rownames = T)

dim(ctrp2_aac)

# dim(ctrp2_ic50)

# Convert to long format

ctrp2_aac <- melt.data.table(data = ctrp2_aac)

# ctrp2_ic50 <- melt.data.table(data = ctrp2_ic50)

# Remove NAs

ctrp2_aac <- ctrp2_aac[!is.na(value)]

# ctrp2_ic50 <- ctrp2_ic50[!is.na(value)]

colnames(ctrp2_aac) <- c("cpd_name", "ccl_name", "area_above_curve")

# colnames(ctrp2_ic50) <- c("cpd_name", "ccl_name", "ic50")



# Add smiles data

dim(ctrp2_aac)  ### 363,634 dose-response curves



dim(ctrp2_ic50)

dim(ctrp2)

ctrp2_smiles <- data.table(drugid = ctrp2@drug$drugid, cpd_smiles = ctrp2@drug$cpd_smiles)



sum(unique(ctrp2@drug$drugid) %in% unique(ctrp2_aac$cpd_name)) / length(unique(ctrp2_aac$cpd_name))





length(unique(ctrp2_aac$cpd_name))

length(unique(ctrp2_ic50$cpd_name))

length(unique(ctrp2_smiles$drugid))



sum(unique(ctrp2_smiles$drugid) %in% unique(ctrp2_aac$cpd_name)) / length(unique(ctrp2_aac$cpd_name))

sum(unique(ctrp2_smiles$drugid) == unique(ctrp2_ic50$cpd_name))



sum(tolower(unique(ctrp2_smiles$drugid)) == tolower(unique(ctrp2_aac$cpd_name)))

sum(tolower(unique(ctrp2_smiles$drugid)) == tolower(unique(ctrp2_ic50$cpd_name)))

changed = !(tolower(unique(ctrp2_smiles$drugid)) == tolower(unique(ctrp2_aac$cpd_name)))

changed = !tolower(unique(ctrp2_smiles$drugid)) == tolower(unique(ctrp2_ic50$cpd_name))



temp = data.table(before = unique(ctrp2_smiles$drugid)[changed], after = unique(ctrp2_aac$cpd_name)[changed])

temp = data.table(before = unique(ctrp2_smiles$drugid)[changed], after = unique(ctrp2_ic50$cpd_name)[changed])

View(temp)



ctrp2_smiles

ctrp2_aac



unique(ctrp2_smiles$drugid)[!(unique(ctrp2_smiles$drugid) == unique(ctrp2_aac$cpd_name))]

unique(ctrp2_aac$cpd_name)[!(unique(ctrp2_smiles$cpd_name) == unique(ctrp2_aac$cpd_name))]



drug_table <- data.table(canonical_name = rownames(drugInfo(ctrp2)),

                         common_name = drugInfo(ctrp2)[, c("cpd_name")],

                         cpd_smiles = drugInfo(ctrp2)[, c("cpd_smiles")])

final_ctrpv2_aac <- merge(ctrp2_aac, drug_table[, c("canonical_name", "cpd_smiles")], by.x = "cpd_name", by.y = "canonical_name")



final_ctrpv2_ic50 <- merge(ctrp2_ic50, drug_table[, c("canonical_name", "cpd_smiles")], by.x = "drugid", by.y = "canonical_name")



# Save

fwrite(final_ctrpv2_aac, paste0(path, "Data/DRP_Training_Data/CTRP_AAC_SMILES.txt"))

fwrite(final_ctrpv2_ic50, paste0(path, "Data/DRP_Training_Data/CTRP_IC50_SMILES.txt"))

final_ctrpv2_aac <- fread(paste0(path, "Data/DRP_Training_Data/CTRP_AAC_SMILES.txt"))

final_ctrpv2_ic50 <- fread(paste0(path, "Data/DRP_Training_Data/CTRP_IC50_SMILES.txt"))



# Add disease information

require(stringr)

line_info <- fread("Data/DRP_Training_Data/DepMap_20Q2_Line_Info.csv")

final_ctrpv2_aac <- fread("Data/DRP_Training_Data/CTRP_AAC_SMILES.txt")



# Remove all hyphens, convert to upper case

line_info$other_ccl_name <- str_replace_all(toupper(line_info$stripped_cell_line_name), "-", "")

final_ctrpv2_aac$other_ccl_name <- str_replace_all(toupper(final_ctrpv2_aac$ccl_name), "-", "")

# Remove all spaces

line_info$other_ccl_name <- str_replace_all(toupper(line_info$other_ccl_name), " ", "")

final_ctrpv2_aac$other_ccl_name <- str_replace_all(toupper(final_ctrpv2_aac$other_ccl_name), " ", "")



sum(unique(final_ctrpv2_aac$other_ccl_name) %in% unique(line_info$other_ccl_name)) / length(unique(final_ctrpv2_aac$other_ccl_name))  ### 0.89177!!!

# A quarter of the data cannot be paired...



# Find what cannot be paired

final_ctrpv2_aac[other_ccl_name %in% unique(final_ctrpv2_aac$other_ccl_name)[!(unique(final_ctrpv2_aac$other_ccl_name) %in% unique(line_info$other_ccl_name))]]



final_ctrpv2_aac <- merge(final_ctrpv2_aac, line_info[, c("other_ccl_name", "primary_disease")], by = "other_ccl_name")

final_ctrpv2_aac$other_ccl_name <- NULL

setcolorder(final_ctrpv2_aac, neworder = c("cpd_name", "ccl_name", "primary_disease", "area_above_curve", "cpd_smiles"))



fwrite(final_ctrpv2_aac, "Data/DRP_Training_Data/CTRP_AAC_SMILES.txt")



# ==== Update August 2021 =====

# Fix Tipifarnbi SMILES

require(data.table)

ctrp_aac <- fread("Data/DRP_Training_Data/CTRP_AAC_SMILES.txt")

ctrp_aac[cpd_name == "Tipifarnib"]$cpd_smiles <- "CN1C=NC=C1C(C2=CC=C(C=C2)Cl)(C3=CC4=C(C=C3)N(C(=O)C=C4C5=CC(=CC=C5)Cl)C)N"

fwrite(ctrp_aac, "Data/DRP_Training_Data/CTRP_AAC_SMILES.txt")



# ctrp_aac <- fread("Data/DRP_Training_Data/CTRP_AAC_SMILES.txt")

# ==== Update September 2021 ====

# Add canonical SMILES for some drugs

require(data.table)

gdsc1_aac <- fread("Data/DRP_Training_Data/GDSC1_AAC_SMILES.txt")

gdsc1_aac[cpd_name == "Tamoxifen"]$cpd_smiles <- "CCC(=C(C1=CC=CC=C1)C2=CC=C(C=C2)OCCN(C)C)C3=CC=CC=C3"

gdsc1_aac[cpd_name == "JW-7-52-1"]$cpd_smiles <- "CCC(=O)N1CCN(CC1)C2=C(C=C(C=C2)N3C(=O)C=CC4=CN=C5C=CC(=CC5=C43)C6=CC7=CC=CC=C7N=C6)C(F)(F)F"

gdsc1_aac[cpd_name == "KIN001-135"]$cpd_smiles <- "COC1=C(C=C2C(=C1)N=CN2C3=CC(=C(S3)C#N)OCC4=CC=CC=C4S(=O)(=O)C)OC"

fwrite(gdsc1_aac, "Data/DRP_Training_Data/GDSC1_AAC_SMILES.txt")



gdsc2_aac <- fread("Data/DRP_Training_Data/GDSC2_AAC_SMILES.txt")

gdsc2_aac[cpd_name == "Tamoxifen"]$cpd_smiles <- "CCC(=C(C1=CC=CC=C1)C2=CC=C(C=C2)OCCN(C)C)C3=CC=CC=C3"

fwrite(gdsc2_aac, "Data/DRP_Training_Data/GDSC2_AAC_SMILES.txt")



# CCLE ====

# ccle <- downloadPSet("CCLE")

# ccle_auc <- summarizeSensitivityProfiles(pSet = ccle, sensitivity.measure = "auc_recomputed", summary.stat = "median", fill.missing = T)

# # Convert to data.table

# ccle_auc <- as.data.table(ccle_auc, keep.rownames = T)

# dim(ccle_auc)

# # Convert to long format

# ccle_auc <- melt.data.table(data = ccle_auc)

# # Remove NAs

# ccle_auc <- ccle_auc[!is.na(value)]

# colnames(ccle_auc) <- c("cpd_name", "ccl_name", "area_under_curve")

# drug_table <- data.table(canonical_name = rownames(drugInfo(ccle)), common_name = drugInfo(ccle)[, c("drug.name")], pubchem = drugInfo(ccle)[, "PubCHEM"])

# ccle_smiles <- webchem::pc_prop(cid = drug_table$pubchem, properties = "CanonicalSMILES")

# # Remove drugs without pubchem entries

# sum(is.na(drug_table$pubchem))  # 15

# drug_table <- drug_table[!is.na(pubchem)]

# drug_table <- merge(drug_table, ccle_smiles, by.x = "pubchem", by.y = "CID")

# 

# 

# ccle_final <- merge(ccle_auc, drug_table[, c("canonical_name", "CanonicalSMILES")], by.x = "cpd_name", by.y = "canonical_name")

# # Save

# fwrite(ccle_final, paste0(path, "Data/DRP_Training_Data/CCLE_AUC_SMILES.txt"))





# GDSC1 ====

gdsc2 <- downloadPSet("GDSC")

gdsc2_auc <- summarizeSensitivityProfiles(pSet = gdsc2, sensitivity.measure = "auc_recomputed", summary.stat = "median", fill.missing = T)

# Convert to data.table

gdsc2_auc <- as.data.table(gdsc2_auc, keep.rownames = T)

dim(gdsc2_auc)

# Convert to long format

gdsc2_auc <- melt.data.table(data = gdsc2_auc)

# Remove NAs

gdsc2_auc <- gdsc2_auc[!is.na(value)]

colnames(gdsc2_auc) <- c("cpd_name", "ccl_name", "area_under_curve")

drug_table <- data.table(canonical_name = rownames(drugInfo(gdsc2)), common_name = drugInfo(gdsc2)[, c("drug.name")], pubchem = drugInfo(gdsc2)[, "PubCHEM"])

gdsc2_smiles <- webchem::pc_prop(cid = drug_table$pubchem, properties = "CanonicalSMILES")

# Remove drugs without pubchem entries

sum(is.na(drug_table$pubchem))  # 15

drug_table <- drug_table[!is.na(pubchem)]

drug_table <- merge(drug_table, gdsc2_smiles, by.x = "pubchem", by.y = "CID")





gdsc2_final <- merge(gdsc2_auc, drug_table[, c("canonical_name", "CanonicalSMILES")], by.x = "cpd_name", by.y = "canonical_name")

# Save

fwrite(gdsc2_final, paste0(path, "Data/DRP_Training_Data/GDSC2_AUC_SMILES.txt"))



gdsc2_final <- fread(paste0(path, "Data/DRP_Training_Data/GDSC2_AUC_SMILES.txt"))

colnames(gdsc2_final)[4] <- "cpd_smiles"