其他
新版TCGA数据库不同癌种的组学数据合并
💡专注R语言在🩺生物医学中的使用
很多文章对于TCGA中的一些癌症都是联合分析的，比如TCGA-COAD和TCGA-READ，首先是它们的疾病特点和治疗方式存在很多相似之处，同时这样做也可以增大样本量。
如果你是使用TCGAbiolinks包下载的数据，那么它们的合并超级简单，直接cbind()即可！
加载数据和R包
数据都是之前下载好的，可以参考之前的推文：
新版TCGAbiolinks包学习：表达矩阵提取（mRNA/lncRNA/counts/tpm/fpkm）
手动下载的TCGA数据也是可以用TCGAbiolinks包整理的
我们直接加载TCGA-COAD和TCGA-READ的数据。
#library(TCGAbiolinks)
# COAD
load(file = "./TCGA-mRNA/TCGA-COAD_mRNA.Rdata")
coad <- data
# READ
load(file = "./TCGA-mRNA/TCGA-READ_mRNA.Rdata")
read <- data
合并数据
现在coad和read都是SummarizedExperiment对象，并且具有相同的行和行名：
coad
## Loading required package: SummarizedExperiment
## Loading required package: MatrixGenerics
## Loading required package: matrixStats
##
## Attaching package: 'MatrixGenerics'
## The following objects are masked from 'package:matrixStats':
##
## colAlls, colAnyNAs, colAnys, colAvgsPerRowSet, colCollapse,
## colCounts, colCummaxs, colCummins, colCumprods, colCumsums,
## colDiffs, colIQRDiffs, colIQRs, colLogSumExps, colMadDiffs,
## colMads, colMaxs, colMeans2, colMedians, colMins, colOrderStats,
## colProds, colQuantiles, colRanges, colRanks, colSdDiffs, colSds,
## colSums2, colTabulates, colVarDiffs, colVars, colWeightedMads,
## colWeightedMeans, colWeightedMedians, colWeightedSds,
## colWeightedVars, rowAlls, rowAnyNAs, rowAnys, rowAvgsPerColSet,
## rowCollapse, rowCounts, rowCummaxs, rowCummins, rowCumprods,
## rowCumsums, rowDiffs, rowIQRDiffs, rowIQRs, rowLogSumExps,
## rowMadDiffs, rowMads, rowMaxs, rowMeans2, rowMedians, rowMins,
## rowOrderStats, rowProds, rowQuantiles, rowRanges, rowRanks,
## rowSdDiffs, rowSds, rowSums2, rowTabulates, rowVarDiffs, rowVars,
## rowWeightedMads, rowWeightedMeans, rowWeightedMedians,
## rowWeightedSds, rowWeightedVars
## Loading required package: GenomicRanges
## Loading required package: stats4
## Loading required package: BiocGenerics
##
## Attaching package: 'BiocGenerics'
## The following objects are masked from 'package:stats':
##
## IQR, mad, sd, var, xtabs
## The following objects are masked from 'package:base':
##
## anyDuplicated, append, as.data.frame, basename, cbind, colnames,
## dirname, do.call, duplicated, eval, evalq, Filter, Find, get, grep,
## grepl, intersect, is.unsorted, lapply, Map, mapply, match, mget,
## order, paste, pmax, pmax.int, pmin, pmin.int, Position, rank,
## rbind, Reduce, rownames, sapply, setdiff, sort, table, tapply,
## union, unique, unsplit, which.max, which.min
## Loading required package: S4Vectors
##
## Attaching package: 'S4Vectors'
## The following objects are masked from 'package:base':
##
## expand.grid, I, unname
## Loading required package: IRanges
##
## Attaching package: 'IRanges'
## The following object is masked from 'package:grDevices':
##
## windows
## Loading required package: GenomeInfoDb
## Loading required package: Biobase
## Welcome to Bioconductor
##
## Vignettes contain introductory material; view with
## 'browseVignettes()'. To cite Bioconductor, see
## 'citation("Biobase")', and for packages 'citation("pkgname")'.
##
## Attaching package: 'Biobase'
## The following object is masked from 'package:MatrixGenerics':
##
## rowMedians
## The following objects are masked from 'package:matrixStats':
##
## anyMissing, rowMedians
## class: RangedSummarizedExperiment
## dim: 60660 521
## metadata(1): data_release
## assays(6): unstranded stranded_first ... fpkm_unstrand fpkm_uq_unstrand
## rownames(60660): ENSG00000000003.15 ENSG00000000005.6 ...
## ENSG00000288674.1 ENSG00000288675.1
## rowData names(10): source type ... hgnc_id havana_gene
## colnames(521): TCGA-A6-5664-01A-21R-1839-07
## TCGA-D5-6530-01A-11R-1723-07 ... TCGA-A6-2683-01A-01R-0821-07
## TCGA-A6-2683-11A-01R-A32Z-07
## colData names(107): barcode patient ... paper_vascular_invasion_present
## paper_vital_status
read
## class: RangedSummarizedExperiment
## dim: 60660 177
## metadata(1): data_release
## assays(6): unstranded stranded_first ... fpkm_unstrand fpkm_uq_unstrand
## rownames(60660): ENSG00000000003.15 ENSG00000000005.6 ...
## ENSG00000288674.1 ENSG00000288675.1
## rowData names(10): source type ... hgnc_id havana_gene
## colnames(177): TCGA-AG-3580-01A-01R-0821-07
## TCGA-AF-2692-11A-01R-A32Z-07 ... TCGA-AG-3894-01A-01R-1119-07
## TCGA-AG-3574-01A-01R-0821-07
## colData names(107): barcode patient ... paper_vascular_invasion_present
## paper_vital_status
对于这样的数据我们直接合并即可，我认为这是目前合并两个癌种最方便的方法了！
# 直接cbind
colrectal <- cbind(coad,read)
colrectal
## class: RangedSummarizedExperiment
## dim: 60660 698
## metadata(2): data_release data_release
## assays(6): unstranded stranded_first ... fpkm_unstrand fpkm_uq_unstrand
## rownames(60660): ENSG00000000003.15 ENSG00000000005.6 ...
## ENSG00000288674.1 ENSG00000288675.1
## rowData names(10): source type ... hgnc_id havana_gene
## colnames(698): TCGA-A6-5664-01A-21R-1839-07
## TCGA-D5-6530-01A-11R-1723-07 ... TCGA-AG-3894-01A-01R-1119-07
## TCGA-AG-3574-01A-01R-0821-07
## colData names(107): barcode patient ... paper_vascular_invasion_present
## paper_vital_status
得到的结果也是一个SummarizedExperiment对象。并且这个对象中各种信息也是保存好的，想用什么直接提取即可，非常方便。
但是这样合并可能涉及批次效应的问题，大家在实际使用时可根据自己的情况选择要不要去除批次效应！
提取信息
比如提取样本的临床信息，非常简单，甚至不需要重新下载：
clin <- as.data.frame(colData(colrectal))
clin[1:10,1:10]
## barcode patient
## TCGA-A6-5664-01A-21R-1839-07 TCGA-A6-5664-01A-21R-1839-07 TCGA-A6-5664
## TCGA-D5-6530-01A-11R-1723-07 TCGA-D5-6530-01A-11R-1723-07 TCGA-D5-6530
## TCGA-AA-3556-01A-01R-0821-07 TCGA-AA-3556-01A-01R-0821-07 TCGA-AA-3556
## TCGA-AA-3660-11A-01R-1723-07 TCGA-AA-3660-11A-01R-1723-07 TCGA-AA-3660
## TCGA-AA-3818-01A-01R-0905-07 TCGA-AA-3818-01A-01R-0905-07 TCGA-AA-3818
## TCGA-AA-3660-01A-01R-1723-07 TCGA-AA-3660-01A-01R-1723-07 TCGA-AA-3660
## TCGA-DM-A28G-01A-11R-A16W-07 TCGA-DM-A28G-01A-11R-A16W-07 TCGA-DM-A28G
## TCGA-AA-3976-01A-01R-1022-07 TCGA-AA-3976-01A-01R-1022-07 TCGA-AA-3976
## TCGA-G4-6307-01A-11R-1723-07 TCGA-G4-6307-01A-11R-1723-07 TCGA-G4-6307
## TCGA-AA-3522-11A-01R-A32Z-07 TCGA-AA-3522-11A-01R-A32Z-07 TCGA-AA-3522
## sample shortLetterCode
## TCGA-A6-5664-01A-21R-1839-07 TCGA-A6-5664-01A TP
## TCGA-D5-6530-01A-11R-1723-07 TCGA-D5-6530-01A TP
## TCGA-AA-3556-01A-01R-0821-07 TCGA-AA-3556-01A TP
## TCGA-AA-3660-11A-01R-1723-07 TCGA-AA-3660-11A NT
## TCGA-AA-3818-01A-01R-0905-07 TCGA-AA-3818-01A TP
## TCGA-AA-3660-01A-01R-1723-07 TCGA-AA-3660-01A TP
## TCGA-DM-A28G-01A-11R-A16W-07 TCGA-DM-A28G-01A TP
## TCGA-AA-3976-01A-01R-1022-07 TCGA-AA-3976-01A TP
## TCGA-G4-6307-01A-11R-1723-07 TCGA-G4-6307-01A TP
## TCGA-AA-3522-11A-01R-A32Z-07 TCGA-AA-3522-11A NT
## definition sample_submitter_id
## TCGA-A6-5664-01A-21R-1839-07 Primary solid Tumor TCGA-A6-5664-01A
## TCGA-D5-6530-01A-11R-1723-07 Primary solid Tumor TCGA-D5-6530-01A
## TCGA-AA-3556-01A-01R-0821-07 Primary solid Tumor TCGA-AA-3556-01A
## TCGA-AA-3660-11A-01R-1723-07 Solid Tissue Normal TCGA-AA-3660-11A
## TCGA-AA-3818-01A-01R-0905-07 Primary solid Tumor TCGA-AA-3818-01A
## TCGA-AA-3660-01A-01R-1723-07 Primary solid Tumor TCGA-AA-3660-01A
## TCGA-DM-A28G-01A-11R-A16W-07 Primary solid Tumor TCGA-DM-A28G-01A
## TCGA-AA-3976-01A-01R-1022-07 Primary solid Tumor TCGA-AA-3976-01A
## TCGA-G4-6307-01A-11R-1723-07 Primary solid Tumor TCGA-G4-6307-01A
## TCGA-AA-3522-11A-01R-A32Z-07 Solid Tissue Normal TCGA-AA-3522-11A
## sample_type_id
## TCGA-A6-5664-01A-21R-1839-07 01
## TCGA-D5-6530-01A-11R-1723-07 01
## TCGA-AA-3556-01A-01R-0821-07 01
## TCGA-AA-3660-11A-01R-1723-07 11
## TCGA-AA-3818-01A-01R-0905-07 01
## TCGA-AA-3660-01A-01R-1723-07 01
## TCGA-DM-A28G-01A-11R-A16W-07 01
## TCGA-AA-3976-01A-01R-1022-07 01
## TCGA-G4-6307-01A-11R-1723-07 01
## TCGA-AA-3522-11A-01R-A32Z-07 11
## sample_id
## TCGA-A6-5664-01A-21R-1839-07 3048539a-b914-4e43-b1cc-43ea707e3b3d
## TCGA-D5-6530-01A-11R-1723-07 50560725-c72d-4bab-b602-5e50e6bececd
## TCGA-AA-3556-01A-01R-0821-07 4794413c-ed92-451c-a3ce-f411fed5ca82
## TCGA-AA-3660-11A-01R-1723-07 a0832917-75b9-45c8-9273-009c3737a43a
## TCGA-AA-3818-01A-01R-0905-07 0cf35153-2c04-4bdd-91e0-d63cb98da5bf
## TCGA-AA-3660-01A-01R-1723-07 87cf1a20-2dc5-4c06-b0c4-16103be40ef0
## TCGA-DM-A28G-01A-11R-A16W-07 f5acd8b8-32c4-4f3c-aacd-259f8e1fdfee
## TCGA-AA-3976-01A-01R-1022-07 8d529023-abca-4ddc-a265-d5bd1fd48708
## TCGA-G4-6307-01A-11R-1723-07 801b8d05-2d29-4f8c-8d8d-634b2e21b867
## TCGA-AA-3522-11A-01R-A32Z-07 218bbd07-5fa3-4946-a2c1-0ece13466441
## sample_type days_to_collection
## TCGA-A6-5664-01A-21R-1839-07 Primary Tumor NA
## TCGA-D5-6530-01A-11R-1723-07 Primary Tumor NA
## TCGA-AA-3556-01A-01R-0821-07 Primary Tumor NA
## TCGA-AA-3660-11A-01R-1723-07 Solid Tissue Normal NA
## TCGA-AA-3818-01A-01R-0905-07 Primary Tumor NA
## TCGA-AA-3660-01A-01R-1723-07 Primary Tumor NA
## TCGA-DM-A28G-01A-11R-A16W-07 Primary Tumor 3419
## TCGA-AA-3976-01A-01R-1022-07 Primary Tumor NA
## TCGA-G4-6307-01A-11R-1723-07 Primary Tumor NA
## TCGA-AA-3522-11A-01R-A32Z-07 Solid Tissue Normal NA
dim(clin)
## [1] 698 107
colnames(clin)[10:30]
## [1] "days_to_collection" "state"
## [3] "initial_weight" "intermediate_dimension"
## [5] "pathology_report_uuid" "submitter_id"
## [7] "shortest_dimension" "oct_embedded"
## [9] "longest_dimension" "is_ffpe"
## [11] "tissue_type" "synchronous_malignancy"
## [13] "ajcc_pathologic_stage" "days_to_diagnosis"
## [15] "treatments" "last_known_disease_status"
## [17] "tissue_or_organ_of_origin" "days_to_last_follow_up"
## [19] "age_at_diagnosis" "primary_diagnosis"
## [21] "prior_malignancy"
现在一共有698行，107列临床信息，你想要的生存时间、生存状态、样本类型、分期等信息都在里面，都不需要自己手动划分，想要什么直接取子集就好了。
比如大家最喜欢的生存信息：
clin_subset <- clin[,c("days_to_last_follow_up","vital_status")]
head(clin_subset)
## days_to_last_follow_up vital_status
## TCGA-A6-5664-01A-21R-1839-07 672 Alive
## TCGA-D5-6530-01A-11R-1723-07 621 Alive
## TCGA-AA-3556-01A-01R-0821-07 700 Alive
## TCGA-AA-3660-11A-01R-1723-07 2375 Alive
## TCGA-AA-3818-01A-01R-0905-07 NA Dead
## TCGA-AA-3660-01A-01R-1723-07 2375 Alive
合并miRNA
也是一样的操作。
rm(list = ls())
load(file = "./TCGA-mirna/TCGA-COAD_miRNA.Rdata")
coad <- data
load(file = "./TCGA-mirna/TCGA-READ_miRNA.Rdata")
read <- data
可以看到两个表达矩阵的第一列（miRNA的名字），完全一样：
identical(coad$miRNA_ID,read$miRNA_ID)
## [1] TRUE
所以我们直接合并即可：
# 第一列都是
colrectal_mi <- cbind(coad,read[,-1])
colrectal_mi[1:5,1:4]
## miRNA_ID read_count_TCGA-A6-5664-01A-21H-1838-13
## 1 hsa-let-7a-1 6959
## 2 hsa-let-7a-2 6941
## 3 hsa-let-7a-3 7120
## 4 hsa-let-7b 31616
## 5 hsa-let-7c 4211
## reads_per_million_miRNA_mapped_TCGA-A6-5664-01A-21H-1838-13
## 1 8143.201
## 2 8122.137
## 3 8331.598
## 4 36996.038
## 5 4927.578
## cross-mapped_TCGA-A6-5664-01A-21H-1838-13
## 1 N
## 2 N
## 3 N
## 4 N
## 5 N
但是miRNA的表达矩阵现在还有点问题，它包含3种信息：count/rpm/cross-mapped，而我们只需要count，所以还是要处理一下。
dim(colrectal_mi)
## [1] 1881 1891
# 只要count
colrec_mi <- colrectal_mi[,c(1,seq(2,1891,by=3))]
dim(colrec_mi)
## [1] 1881 631
# 改下列名
colnames(colrec_mi)[-1] <- substr(colnames(colrec_mi)[-1],12,39)
colrec_mi[1:5,1:5]
## miRNA_ID TCGA-A6-5664-01A-21H-1838-13 TCGA-A6-2683-01A-01T-0822-13
## 1 hsa-let-7a-1 6959 50288
## 2 hsa-let-7a-2 6941 50537
## 3 hsa-let-7a-3 7120 51098
## 4 hsa-let-7b 31616 143822
## 5 hsa-let-7c 4211 3943
## TCGA-D5-6530-01A-11H-1722-13 TCGA-DM-A28G-01A-11H-A16S-13
## 1 35778 11788
## 2 35334 11588
## 3 35980 11885
## 4 68674 12086
## 5 605 1171
简单！
合并CNV
rm(list = ls())
load("G:/tcga/TCGA-CNV/TCGA-COAD_CNV.Rdata")
coad <- data
load("G:/tcga/TCGA-CNV/TCGA-READ_CNV.Rdata")
read <- data
colrec_cnv <- rbind(coad,read)
head(colrec_cnv)
## GDC_Aliquot Chromosome Start End Num_Probes
## 1 741d4882-3a2c-4862-8402-636e6aebfdc6 1 3301765 247650984 129758
## 2 741d4882-3a2c-4862-8402-636e6aebfdc6 2 480597 241537572 132218
## 3 741d4882-3a2c-4862-8402-636e6aebfdc6 3 2170634 25586863 14093
## 4 741d4882-3a2c-4862-8402-636e6aebfdc6 3 25587626 25587698 3
## 5 741d4882-3a2c-4862-8402-636e6aebfdc6 3 25588064 197812401 93106
## 6 741d4882-3a2c-4862-8402-636e6aebfdc6 4 1059384 124538250 69561
## Segment_Mean Sample
## 1 -0.0019 TCGA-AA-3556-10A-01D-0819-01
## 2 -0.0007 TCGA-AA-3556-10A-01D-0819-01
## 3 -0.0009 TCGA-AA-3556-10A-01D-0819-01
## 4 -1.9166 TCGA-AA-3556-10A-01D-0819-01
## 5 0.0023 TCGA-AA-3556-10A-01D-0819-01
## 6 0.0025 TCGA-AA-3556-10A-01D-0819-01
这个文件稍加整理就可以拿去给gistic用了。
合并SNP
rm(list = ls())
load("G:/tcga/TCGA-SNP/TCGA-READ_SNP.Rdata")
read <- data
load("G:/tcga/TCGA-SNP/TCGA-COAD_SNP.Rdata")
coad <- data
colrec_snp <- rbind(coad,read)
这样以后再分析就可以用合并后的数据了！
获取更多信息，欢迎加入🐧QQ交流群：613637742
“医学和生信笔记，专注R语言在临床医学中的使用、R语言数据分析和可视化。主要分享R语言做医学统计学、meta分析、网络药理学、临床预测模型、机器学习、生物信息学等。
往期回顾
tidymodels用于机器学习的一些使用细节
生信初学者基础知识资源推荐
比例韦恩图
初学者使用R语言读取、写出文件(csv/txt/excel/rdata等)的注意事项
R语言和医学统计学（合辑）