LINCS1与LINCS2数据库

https://docs.google.com/document/d/1q2gciWRhVCAAnlvF2iRLuJ7whrGP6QjpsCMq1yWz7dU

Part 1、LINCS Phase I L1000--GSE92742

• https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE92742

1、signature矩阵

library(cmapR)
library(tidyverse)

gctx_demo = parse_gctx("GSE92742_Broad_LINCS_Level5_COMPZ.MODZ_n473647x12328.gctx", 
                       cid=1:4, rid=1:4)
gctx_demo@mat
     # CPC005_A375_6H:BRD-A85280935-003-01-7:10
# 5720                                0.7737690
# 466                                -0.8184680
# 6009                                0.1895723
# 2309                               -0.1460308
     # CPC005_A375_6H:BRD-A07824748-001-02-6:10
# 5720                               -0.6455861
# 466                                -0.8107487
# 6009                                0.4590603
# 2309                               -0.2246765
     # CPC004_A375_6H:BRD-K20482099-001-01-1:10
# 5720                                -5.449666
# 466                                  2.393775
# 6009                                 1.279790
# 2309                                 2.167868
     # CPC005_A375_6H:BRD-K62929068-001-03-3:10
# 5720                                0.1934077
# 466                                -0.5822433
# 6009                               -0.1789770
# 2309                               -1.1820246

2、signature注释信息

• 从GEO下载的顺序与上面gct矩阵的sig顺序不一致，需要调整与gct保持一致，方便后续取子集操作。
  如下是已经调整好的：

sig_info = data.table::fread("Fine_phase1_sig_info_473647.csv",
                             data.table = F)

col_meta <- read_gctx_meta("GSE92742_Broad_LINCS_Level5_COMPZ.MODZ_n473647x12328.gctx", dim="col")
identical(col_meta$id, sig_info$sig_id)
# [1] TRUE

dim(sig_info)
# [1] 473647     12
t(sig_info[1,])
#                 [,1]                                                                                                            
# sig_id         "CPC005_A375_6H:BRD-A85280935-003-01-7:10"                                                                      
# pert_id        "BRD-A85280935"
# pert_iname     "quinpirole"                                                                                                    
# pert_type      "trt_cp"                                                                                                        
# cell_id        "A375"                                                                                                          
# pert_dose      "10.0"                                                                                                          
# pert_dose_unit "µM"                                                                                                            
# pert_idose     "10 µM"                                                                                                         
# pert_time      "6"                                                                                                             
# pert_time_unit "h"                                                                                                             
# pert_itime     "6 h"                                                                                                           
# distil_id      "CPC005_A375_6H_X1_B3_DUO52HI53LO:K06|CPC005_A375_6H_X2_B3_DUO52HI53LO:K06|CPC005_A375_6H_X3_B3_DUO52HI53LO:K06"

3、gene注释信息

• 同上也需要调整顺序

gene_info = data.table::fread("Fine_phase1_gene_info_12328.csv",
                              data.table = F)

row_meta <- read_gctx_meta("GSE92742_Broad_LINCS_Level5_COMPZ.MODZ_n473647x12328.gctx", dim="row")
identical(as.integer(row_meta$id), gene_info$pr_gene_id)
# [1] TRUE

dim(gene_info)
# [1] 12328     5
t(gene_info[1,])
#                 1                             
# pr_gene_id     "5720"                          
# pr_gene_symbol "PSME1"                         
# pr_gene_title  "proteasome activator subunit 1"
# pr_is_lm       "1"        # landmark    978                         
# pr_is_bing     "1"        # landmark + best inferred gene   10174

4、pertubation(化合物)注释信息

• 可结合signature注释信息，做进一步筛选

pert_info = data.table::fread("phase1_pert_info.csv")
t(pert_info[1,])
#                   [,1]    
# V1               "1"     
# pert_id          "56582" 
# pert_iname       "AKT2"  
# pert_type        "trt_oe"
# is_touchstone    "0"     
# inchi_key_prefix "-666"  
# inchi_key        "-666"  
# canonical_smiles "-666"  
# pubchem_cid      "-666"  

table(pert_info$pert_type) %>% sort(decreasing = T)
# trt_cp          trt_sh      trt_sh.cgs      trt_sh.css 
# 20413           18493            4345            3807 
# trt_oe         trt_lig      trt_oe.mut      ctl_vector 
# 3492             622             135              61 
#...

5、细胞系注释信息

• 可结合signature注释信息，做进一步筛选

cell_info = data.table::fread("phase1_cell_info.csv")
t(cell_info[1,])
#                         [,1]                
# cell_id                 "A375"              
# cell_type               "cell line"         
# base_cell_id            "A375"              
# precursor_cell_id       "-666"              
# modification            "-666"              
# sample_type             "tumor"             
# primary_site            "skin"              
# subtype                 "malignant melanoma"
# original_growth_pattern "adherent"          
# provider_catalog_id     "CRL-1619"          
# original_source_vendor  "ATCC"              
# donor_age               "54"                
# donor_sex               "F"                 
# donor_ethnicity         "-666"  

table(cell_info$sample_type)
# -666  normal primary   tumor 
# 1      19       8      70

6、signature干扰效应评价

• Replicate Correlation Coefficient
• signature strength
• Transcriptional Activity Score

sig_metrics = data.table::fread("phase1_sig_metrcs.csv")

identical(sig_metrics$sig_id, col_meta$id)
# [1] TRUE

t(sig_metrcs[1,])
# [,1]
# sig_id                 "CPC005_A375_6H:BRD-A85280935-003-01-7:10"
# pert_id                "BRD-A85280935"
# pert_iname             "quinpirole"
# pert_type              "trt_cp"
# distil_cc_q75          "0.11"   (Replicate Correlation Coefficient)           
# distil_ss              "2.84895" (signature strength)    
# ngenes_modulated_up_lm "18"
# ngenes_modulated_dn_lm "15"
# tas                    "0.101169" (Transcriptional Activity Score)        
# pct_self_rank_q25      "7.6087"
# is_exemplar            "0"
# distil_nsample         "3"

Part 2、LINCS Phase II L1000--GSE70138

• https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE70138
• 收集、整理数据步骤基本同上，不赘述了。