R-msigdf + clusterProfiler全方位支持MSigDb

生信界的网红Stephen Turner在github上有个msigdf的包，我在他写这个包的时候，就写了个gist，连接clusterProfiler，我写gist的时候是2016年的8月，很高兴网红还惦记着我的gist。

msigdf这个包把著名的Broad Institute著名的Molecular Signatures Database (MSigDB)数据以data frame的形式打包成R包，这样子非常方便使用，当然他后来没有更新，而一个fork的版本，ToledoEM/msigdf把数据更新为最新版本v6.2，发布于2018年7月。

这个包，天生就方便我们clusterProfiler用户，我在《Comparison of clusterProfiler and GSEA-P》一文中，为了比较clusterProfiler和Broad Institute出品的GSEA-P软件，特意打包了一个gmt文件，这样方便在注释一样的情况下比较，有了这个示例，其实大家就应该知道，下载gmt文件，然后就可以用clusterProfiler分析了，也就是说clusterProfiler是支持MSigDb的，再一次敲重点，不要再以为clusterProfiler只做GO和KEGG了，clusterProfiler啥都能干。

好了，那么有这个包之后，我们连下载gmt文件都省了，直接载入这个包的数据，然后就可以无缝衔接clusterProfiler进行分析，还是原来的配方，还是熟悉的味道，你拥有了《enrichplot: 让你们对clusterProfiler系列包无法自拔》的各种可视化功能。

`msigdf`包

## devtools::install_github("ToledoEM/msigdf")
library(msigdf)

msigdf包含了人和鼠的数据，总共有以下8个类别：

category	description
H	hallmark gene sets
C1	positional gene sets
C2	curated gene sets
C3	motif gene sets
C4	computational gene sets
C5	GO gene sets
C6	oncogenic signatures
C7	immunologic signatures

在这里我将过滤出C2拿来做分析：

library(dplyr)

c2 <- msigdf.human %>% 
    filter(category_code == "c2") %>% select(geneset, symbol) %>% as.data.frame

下面这就是我们拿到的C2的注释，一个data.frame：

> head(c2, 3)
                              geneset  symbol
1 NAKAMURA_CANCER_MICROENVIRONMENT_UP  COL1A2
2 NAKAMURA_CANCER_MICROENVIRONMENT_UP GPIHBP1
3 NAKAMURA_CANCER_MICROENVIRONMENT_UP     RET

接下来是clusterProfiler的表演时间，我们还是使用DOSE包里的geneList数据来做演示，如果不知道怎么搞自己的geneList，请移步《听说你有RNAseq数据却不知道怎么跑GSEA》。

library(clusterProfiler)
data(geneList, package="DOSE")

这里涉及到一个问题，geneList里的基因是ENTREZID，而这里msigdf包里是SYMBOL，这个简单，无非是个基因ID转换的过程而已，一般而言，对于你自己的注释数据，你不应当有这个问题；假如有，也是一个转换步骤而已，这里做为演示，我把geneList的ID给转了，你当然也可以转注释c2里的ID。

> id = bitr(names(geneList), "ENTREZID", "SYMBOL", "org.Hs.eg.db")
'select()' returned 1:1 mapping between keys and columns
Warning message:
In bitr(names(geneList), "ENTREZID", "SYMBOL", "org.Hs.eg.db") :
  0.41% of input gene IDs are fail to map...
> head(id, 3)
  ENTREZID SYMBOL
1     4312   MMP1
2     8318  CDC45
3    10874    NMU

我们看到bitr的输出，有部分ID是没法转的，这些应该过滤掉，然后用新的ID去给geneList重命名即可，要注意ID要一一对应，不能搞错，我这里用了match，确保操作没问题：

> geneList = geneList[names(geneList) %in% id[,1]]
> names(geneList) = id[match(names(geneList), id[,1]), 2]

好了，现在轮到熟悉的clusterProfiler分析了，大家应该都不陌生，在下面的文章中都有介绍过：

用enricher做超几何分布检验：

de <- names(geneList)[1:100]
x <- enricher(de, TERM2GENE = c2)

用head瞄一眼结果：

> head(x, 3)
                                                                                     ID
ROSTY_CERVICAL_CANCER_PROLIFERATION_CLUSTER ROSTY_CERVICAL_CANCER_PROLIFERATION_CLUSTER
SOTIRIOU_BREAST_CANCER_GRADE_1_VS_3_UP           SOTIRIOU_BREAST_CANCER_GRADE_1_VS_3_UP
SHEDDEN_LUNG_CANCER_POOR_SURVIVAL_A6               SHEDDEN_LUNG_CANCER_POOR_SURVIVAL_A6
                                                                            Description
ROSTY_CERVICAL_CANCER_PROLIFERATION_CLUSTER ROSTY_CERVICAL_CANCER_PROLIFERATION_CLUSTER
SOTIRIOU_BREAST_CANCER_GRADE_1_VS_3_UP           SOTIRIOU_BREAST_CANCER_GRADE_1_VS_3_UP
SHEDDEN_LUNG_CANCER_POOR_SURVIVAL_A6               SHEDDEN_LUNG_CANCER_POOR_SURVIVAL_A6
                                            GeneRatio   BgRatio       pvalue
ROSTY_CERVICAL_CANCER_PROLIFERATION_CLUSTER     49/99 140/21142 5.021928e-83
SOTIRIOU_BREAST_CANCER_GRADE_1_VS_3_UP          46/99 151/21142 2.834141e-74
SHEDDEN_LUNG_CANCER_POOR_SURVIVAL_A6            58/99 456/21142 3.097882e-71
                                                p.adjust       qvalue
ROSTY_CERVICAL_CANCER_PROLIFERATION_CLUSTER 8.050151e-80 5.370820e-80
SOTIRIOU_BREAST_CANCER_GRADE_1_VS_3_UP      2.271564e-71 1.515520e-71
SHEDDEN_LUNG_CANCER_POOR_SURVIVAL_A6        1.655302e-68 1.104368e-68
                                                                                                                                                                                                                                                                                                                                                                                            geneID
ROSTY_CERVICAL_CANCER_PROLIFERATION_CLUSTER                                                   CDCA8/MCM10/CDC20/FOXM1/KIF23/CENPE/MYBL2/MELK/CCNB2/NDC80/TOP2A/NCAPH/E2F8/ASPM/RRM2/CEP55/DLGAP5/UBE2C/HJURP/APOBEC3B/PBK/NUSAP1/CDCA3/TPX2/TACC3/NEK2/CENPM/RAD51AP1/UBE2S/CCNA2/CDK1/ERCC6L/MAD2L1/GINS1/BIRC5/KIF11/EZH2/TTK/NCAPG/AURKB/CHAF1B/CHEK1/TRIP13/PRC1/KIFC1/KIF18B/KIF20A/DTL/AURKA
SOTIRIOU_BREAST_CANCER_GRADE_1_VS_3_UP                                                                           CDC45/CDCA8/MCM10/CDC20/FOXM1/CENPE/MYBL2/MELK/CCNB2/NDC80/TOP2A/NCAPH/E2F8/ASPM/RRM2/CEP55/DLGAP5/UBE2C/HJURP/APOBEC3B/NUSAP1/CDCA3/TPX2/TACC3/NEK2/SLC7A5/CENPN/UBE2S/CCNA2/CDK1/MAD2L1/GINS1/BIRC5/KIF11/EZH2/TTK/NCAPG/AURKB/CHEK1/TRIP13/PRC1/KIFC1/KIF18B/QPRT/KIF20A/AURKA
SHEDDEN_LUNG_CANCER_POOR_SURVIVAL_A6        CDC45/NMU/CDCA8/MCM10/CDC20/FOXM1/KIF23/CENPE/MYBL2/MELK/CCNB2/NDC80/TOP2A/NCAPH/E2F8/ASPM/RRM2/CEP55/DLGAP5/UGT8/UBE2C/HJURP/APOBEC3B/SKA1/PBK/NUSAP1/CDCA3/TPX2/TACC3/NEK2/SLC7A5/CENPM/RAD51AP1/CENPN/UBE2S/CCNA2/CDK1/ERCC6L/MAD2L1/GINS1/KIF18A/CDT1/BIRC5/KIF11/EZH2/TTK/NCAPG/GPR19/AURKB/GINS2/CHEK1/TRIP13/PRC1/KIF18B/MMP12/KIF20A/DTL/AURKA
                                            Count
ROSTY_CERVICAL_CANCER_PROLIFERATION_CLUSTER    49
SOTIRIOU_BREAST_CANCER_GRADE_1_VS_3_UP         46
SHEDDEN_LUNG_CANCER_POOR_SURVIVAL_A6           58

然后GSEA分析用GSEA函数：

y <- GSEA(geneList, TERM2GENE = c2)

同样我们也用head瞄一眼结果：

> head(y, 3)
                                                                                                     ID
ONKEN_UVEAL_MELANOMA_DN                                                         ONKEN_UVEAL_MELANOMA_DN
NAKAMURA_TUMOR_ZONE_PERIPHERAL_VS_CENTRAL_DN               NAKAMURA_TUMOR_ZONE_PERIPHERAL_VS_CENTRAL_DN
BONOME_OVARIAN_CANCER_SURVIVAL_SUBOPTIMAL_DEBULKING BONOME_OVARIAN_CANCER_SURVIVAL_SUBOPTIMAL_DEBULKING
                                                                                            Description
ONKEN_UVEAL_MELANOMA_DN                                                         ONKEN_UVEAL_MELANOMA_DN
NAKAMURA_TUMOR_ZONE_PERIPHERAL_VS_CENTRAL_DN               NAKAMURA_TUMOR_ZONE_PERIPHERAL_VS_CENTRAL_DN
BONOME_OVARIAN_CANCER_SURVIVAL_SUBOPTIMAL_DEBULKING BONOME_OVARIAN_CANCER_SURVIVAL_SUBOPTIMAL_DEBULKING
                                                    setSize enrichmentScore
ONKEN_UVEAL_MELANOMA_DN                                 491      -0.3915581
NAKAMURA_TUMOR_ZONE_PERIPHERAL_VS_CENTRAL_DN            466      -0.3459209
BONOME_OVARIAN_CANCER_SURVIVAL_SUBOPTIMAL_DEBULKING     479      -0.3882366
                                                          NES      pvalue
ONKEN_UVEAL_MELANOMA_DN                             -1.752063 0.001221001
NAKAMURA_TUMOR_ZONE_PERIPHERAL_VS_CENTRAL_DN        -1.542803 0.001230012
BONOME_OVARIAN_CANCER_SURVIVAL_SUBOPTIMAL_DEBULKING -1.734615 0.001231527
                                                      p.adjust    qvalues rank
ONKEN_UVEAL_MELANOMA_DN                             0.01896267 0.01308232 3641
NAKAMURA_TUMOR_ZONE_PERIPHERAL_VS_CENTRAL_DN        0.01896267 0.01308232 2573
BONOME_OVARIAN_CANCER_SURVIVAL_SUBOPTIMAL_DEBULKING 0.01896267 0.01308232 2967
                                                                      leading_edge
ONKEN_UVEAL_MELANOMA_DN                             tags=41%, list=29%, signal=30%
NAKAMURA_TUMOR_ZONE_PERIPHERAL_VS_CENTRAL_DN        tags=33%, list=21%, signal=27%
BONOME_OVARIAN_CANCER_SURVIVAL_SUBOPTIMAL_DEBULKING tags=37%, list=24%, signal=29%
                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                     core_enrichment
ONKEN_UVEAL_MELANOMA_DN                             CTNNBIP1/DPY19L2P2/CNN3/ADCY6/TBX2/RPL11/KAT2B/C6orf48/SCAP/ETV1/HADHA/EIF3G/EIF1B/HFE/FAM184A/KDM6A/SPIN1/MBNL2/ABHD6/MID2/CSDE1/PRKAR1A/SLC35D2/BAG5/CSGALNACT1/TGOLN2/SOX9/NCKIPSD/ZNF185/QARS/DPYSL2/GABARAPL1/LRRC1/PCNP/SEC62/PEX6/TOP2B/ALDH9A1/FHL2/DLC1/RPL15/ENPP2/ZC3H13/USP4/ATXN7/ZSCAN12/EDNRB/SLC25A36/TWF1/ERBB3/MGEA5/OSBPL10/DSTN/SETD2/MFAP2/TMEM47/FAM86B1/NKTR/TFAP2A/TJP1/ADD3/NFE2L1/ECSIT/DAZAP2/MANSC1/RPL14/ACAA1/SNCA/SEMA6A/RPL29/CD200/TDRD3/APPL1/SDC4/ADRB2/GPR153/CTDSP2/VAMP3/GALC/RSL1D1/PREPL/PBX1/IP6K1/CTNNB1/GMPR2/PDHB/KANK2/DAB2/DBP/CPS1/SERPINB6/ENTPD1/GORASP1/PCLO/PTP4A2/PIK3R1/GLUL/RYBP/LETMD1/SCAMP1/PRCP/LAMA4/NEDD9/SPRY2/XPC/ASAH1/NEK4/GAS1/HBB/EIF4B/IMPDH2/PEG10/TSPYL1/CRBN/GOLGA4/LMCD1/MAGEH1/FSTL3/NRIP1/CCND1/HIST1H2AC/TOM1L1/NDN/GLCE/PALLD/BEX4/CREBL2/ITM2A/SPRY1/ZMAT3/CCDC28A/SOBP/KCNS3/SSBP2/ZBTB38/HSD17B8/ID4/ALDH6A1/IL11RA/C3orf14/CTSF/VWA5A/EFEMP2/RAB17/SNAP23/TIMP3/UBL3/SORBS1/DALRD3/SNRK/HSPA2/ZNF91/JAM3/ARMCX2/HNMT/GSTM2/SERPINI1/FAM129A/CADPS2/LPAR6/PDGFC/GSTM1/TXNIP/CCNG2/TRAK1/MTUS1/CBX7/SMARCA1/MPPED2/CAV1/CIRBP/NAP1L2/DYNC2H1/MZT2B/HEMK1/FYCO1/GPD1L/SLC7A8/PMP22/ZNF415/SLC39A6/POSTN/SYBU/GNG11/ZSCAN18/ZBTB20/LAMB2/LZTFL1/PLSCR4/CARTPT/GSTM3/AZGP1/PDGFD/ZNF423/RNASE4/KCNE4/PSD3/EMCN/PPP1R3C
NAKAMURA_TUMOR_ZONE_PERIPHERAL_VS_CENTRAL_DN                                                                                                                                                                                                                                                                                                                          ETS2/ADD3/SEC16A/ABCC3/TIA1/MACF1/RAB40B/DENND4C/C1RL/MXI1/SEMA6A/HBP1/WDR11/IGSF3/ZNF24/SERPINB1/GALC/RSL1D1/DPYD/FBXO9/NRN1/WSB1/ZNF638/ARNT2/AES/GATAD1/CPS1/NFATC2IP/MAVS/TRA2A/TTC31/BPTF/IGBP1/DTWD1/SRSF5/SLC2A10/LOXL2/STAT6/GLUL/RYBP/ITM2B/LETMD1/BNIP3L/EIF4A1/PHF3/COL6A1/CAST/WFDC2/BCL6/LMO2/IL1R1/AUH/INSR/TESC/GAS1/TAF9B/HP/TNFSF10/DOK1/PIGV/PCDH7/SGSM2/CLEC2B/DUSP1/FUCA1/USP34/HSD17B11/SLC24A3/SNCAIP/AHR/MKNK2/RNF125/FZD10/ZNF451/HIST1H2AC/FGFR3/KLHL28/C2orf68/CFI/NR4A2/ARID5B/GOLGA8A/ASPH/ZNF44/ZNF226/BAG1/ATG14/NELL2/UNC5B/HEY1/ENOSF1/ORAI3/RCOR3/SEZ6L2/NBR1/FNDC3A/SH3BGRL/LUM/FOS/CITED2/SLC46A3/TSPAN1/SPOCK1/CCNG1/DCAF10/RAB31/RAB27B/SPG11/ZNF91/IL33/UGCG/RGL2/CLK4/HNMT/SORL1/LPAR6/ST6GALNAC2/AHNAK2/TXNIP/CCNG2/CFH/LRIG1/ZNF580/CLK1/TPBG/MEGF6/MZF1/SLC7A2/ZNF395/ATP8B1/TTC28/CPE/CCNL2/KDM4B/GRAMD1C/C16orf45/CHST15/N4BP2L2/NISCH/IKBKB/NOTCH2NL/PILRB/PLAT/PODNL1/MAOA/CROT/COL4A5/MUC1/SLC1A1/SYT17/STC2/STEAP4/TMC5/CYBRD1
BONOME_OVARIAN_CANCER_SURVIVAL_SUBOPTIMAL_DEBULKING                                                                                                                                                            STOM/SDC2/STAT5B/ENPP2/FRMD4A/PTPRD/SCAPER/DNAJB4/OBSL1/AVPR2/PTPRS/TCF15/PEX3/ASPSCR1/NR2F1/SMARCA2/MORC3/TTC37/SIRT1/SEMA3F/TLE3/SRSF11/CCNH/NR2F2/SETD2/TMEM47/TNKS/CDK14/BCAT2/ADCY9/LRRC8E/CRTC3/AKAP10/RANBP3/GSTM5/PPP3CB/PDE4A/TM2D1/RECK/SPOP/PCGF3/RASA1/ECE1/BTF3/SPRED2/RAD17/KCNQ1/KCNAB1/ERLIN2/RB1/MRPS27/PPP3CA/EXOC7/METTL3/WBP4/HDAC4/EPB41L3/AP1G2/FEZ1/HOXD3/TMEM168/CCPG1/EID1/PIK3R1/CREBBP/AMH/MAN2A1/ITM2B/BNIP3L/SCAMP1/NID1/UBE2B/KDR/LIMCH1/ASAH1/NHLRC2/STK39/WNT5A/GPR39/FBXO38/FILIP1L/WFS1/ACSL3/PDS5B/LINC00260/HOOK2/ZNF235/F2R/PSD4/GULP1/COL15A1/SLC24A3/APC/SNCAIP/GATM/SFI1/CAMK2N1/NDN/PTPRN2/ARHGEF10/FBXL5/ITM2A/PDCD6/ZNF444/MEF2C/RPS6KA2/RARRES2/LPAR1/EDEM3/FNDC3A/DCHS1/ZFYVE16/LEPR/REEP5/THSD7A/RUNX1T1/ZNF839/COL21A1/UBL3/JAM2/SNAI2/GIPC2/PJA2/SPOCK1/CDH13/SOCS2/SYNE1/NBL1/MYOF/JAM3/PNISR/PDE1A/TRPC1/KL/ANKRA2/FAM172A/FAT4/PTGER3/EFEMP1/TSPAN7/IRAK3/SESN1/CAV1/MKL2/SEMA3C/EDNRA/ZBTB16/TGFB1I1/SPATA7/NBR2/SPDEF/CCDC106/ECM2/ABCG2/MEOX2/IKBKB/AGTR1/ZEB1/GPRASP1/DKK2/IGFBP4/SPARCL1/IRS1/AMIGO2/LMOD1/DCN/MAOA/IL6ST/ZCCHC24/PDZRN3/PDGFD/MAOB/RNASE4/C7/PSD3/CILP/ABCA8

用clusterProfiler分析有多种好处，一个是你可以用compareCluster比较不同的条件，还有就是有非常好的可视化函数，《enrichplot: 让你们对clusterProfiler系列包无法自拔》这篇文章你不容错过。

比如：